Dr. Boghossian, It’s unclear to me what ‘AI Will End Academia’ means. I’m - presently - unimpressed with AI. Seven examples: 1. My university-level philosophy courses required interaction with professors and students - often employing the Socratic Method. The inability to do research without AI assistance would be apparent. 2. Many of my law classes employed the Socratic Method. 3. In law school many final exams are based upon questions / cases never covered. It’s to see if the student understands how to correlate what’s been learned. 4. In a recent case, two lawyers provided information to the court. The information / references used in their argument(s) were generated from AI. They were disbarred. 5. I’ve been a Commercial Pilot and Flight Instructor for decades. It takes me seconds to formulate open ended questions stumping AI - questions which don’t stump well trained students. Examples: What is the minimum medical certificate required for an FAA certificated flight instructor? Why? (To date, Dr. B., AI cannot answer this simple question.) 6. When I renew my instructor certificates (every two years), I do this in a classroom setting surrounded other instructors and generally taught by a Ph.D. Why? ‘Learning / Teaching is an active process.’ 7. What would happen if I asked AI, “What does Dr. Peter Boghossian mean when he says ‘AI Will End Academia’?” Will I get the correct answer - one you’ll sign off on? Maybe better to listen to the video and, on occasion, interact with you? :) yYM
You are right, and AI has many more faults on top. Yet, as an academic myself who 'plays a lot' with AI, I found pretty good use cases for it. For example, I was writing a European research grant proposal where I was forced to write in extensive newspeak passages about how my research advances EU priorities, etc. I uploaded a 100-page pdf full of EU political newspeak and the science part of the grant and instructed the AI to find matches and rationalize them. AI did brilliant job at that, saved my time, and lot of anger I always get from reading half-witted bureaucratic nonsense.
Just for fun I put that prompt into ChatGPT. If you happen to read this comment Dr. Boghossian, I'd be curious if you'd give your opinion, perhaps on a scale from Strongly Disagree to Agree, on how much ChatGPT's response reflects your real beliefs. When Dr. Peter Boghossian says "AI will end academia," he refers to a belief that AI will disrupt the current academic system, which he sees as being overrun by ideological conformity and a lack of rigorous critical thinking. Boghossian has been critical of modern universities, which he argues have become dominated by "woke" ideologies that stifle debate and free inquiry. He feels that academia is more focused on promoting certain ideological positions rather than fostering open dialogue and the search for truth. AI, in his view, could further expose and challenge the deficiencies of current academic practices, especially in areas like peer review, where low-quality or ideologically-driven scholarship sometimes gets through. He’s suggested that the automation and analytical capabilities of AI might replace many traditional academic roles or bypass universities altogether, potentially leading to their decline. By contrast, Boghossian supports educational environments that prioritize open discourse, critical thinking, and diverse viewpoints, which he sees as lacking in modern universities.
@@mitzee8621 This is incredible. AI is getting better at an alarming rate. Even a couple of years ago it was so easy to spot AI written articles, now it's much more difficult, and soon it will be impossible.
@@miramichi30 It's already well beyond my ability to spot, with the exception of some technical topics I know a lot about. I don't have much of an eye or ear for AI generated images and music either. If I don't see the person actually creating it in front of me I always harbour doubts as to it's authenticity at this point.
The fact that universities are corrupted institutions that do not punish plagiarists, etc., is actually not a reason to jump further down that hole. I don't want to live in a world where every other engineer and doctor has cheated their way through school using AI and is fully unfamiliar with the fundamentals of their field, sorry. That shit might be fine for the English department but some disciplines actually matter for human lives.
The problem is no different than a factory worker. If you can build a machine that can do the job flawlessly and faster than the best human worker then why do we need them? The factory workers need to learn to build or maintain the machines which destroys one job and creates a different new job for people to do. The only time cheating hurts is when someone can cheat to get a piece of paper that says they can do something when they can't really perform the job. If a computer can do a job better than the human and that computer can be used to replace the real human in the real job then there is no reason for the human to learn to do the job in college that they will never use in the real job. Now this isn't normally a end all learning of one thing because computers can do it now but the requirement to do that part in college needs to be slowly reduced so instead of spending 100 hours learning it they only spend 20 hours learning it so they still have some knowledge of it as a failsafe but knowing they will likely never need to know it in their job but still have enough knowledge to deal with it for the rare case it is needed.
As long as a deep knowledge of a subject is needed, then a way to check that everyone really understands their subject is essential. Any student who submits a thesis, study or any other test should be asked questions about it to determine if they understand what they submitted. For example, "This paragraph you wrote, what do think the main point of it was?" How does that tie in with what you wrote here?" Those sort of questions will quickly determine if they cheated or not. I think AI is great as a practical tool, but not for a way to determine understanding of a subject. For example, I use chatgpt to run through some ideas for inventions that I have thought about.
@@TheSimCaptain I agree there should be a way to see if they really know the stuff. However, in today's world everyone is sitting there with a smart phone and access to the internet. How about to be more realistic allow them to search the internet for an answer to any question you ask them. Think how much more realistic it would be for smart phones and internet access to be allowed when taking a test. As a software dev I don't have every single coding language memorized but I know how to search and find the information I need to write the code. I rarely touch C++ but I have a few times used it to write hacks for games I play to get past long boring parts so I can do other things. I think the ability to look up and find the information you need is the most important skill for almost any job. Not everyone needs to memorize everything but they need to understand the basics and know how to find the information to complete the job.
Um, doctors and engineers have already been taken over by cheaters. Why do you think we have had the feminist bridges that collapsed? All the migrants that was claimed to be nurses and doctors, but without basic western understanding of medicine. All of them bowing to the ideas of DEI. Acknowledging their privilege of their skin color. Non of these institutions are not already corrupt.
It's mostly an issue of assessment validity. All these tools are producing the type of output that students are assessed on. Change your tools of assessment to those which can't be done by AI (personal interviews, in-class assignments...) and you'll be just fine - your degree will be a reliable indicator of competence.
Why do you assume he has built it? Not just plagiarized somebody else work? Or been given it by one of the 3 letter agencies, who do collect user info, to be used for a later date blackmail.
That assumes that people only learn things because of academic requirements. There are plenty of people who study and learn things simply because it interests them.
I have learned the most in papers I've researched and classes I've taught. "Memorization game" is the opposite of this - he's undermining his own point. I feel like stealth-ai wrote all his responses. Also, please explain how science degrees (especially) are worthless.
This is not good. This guy is a straight up villain. Engineers don't need more time to fuck around, they need to study and understand things in a deep technical sense, not pass the work off to an AI.
So he is against the current system of grading where people strive for an A+ but is okay with his AI product for "learning" and grading in which the student partied all night and plugged in his product and in 10 minutes gets a paper for an A. Oh brother! I can see his point if the student does not value the class and is only going through the motions because it is a requirement for a degree so the system would need to be changed. However, it still allows less talented and/or hard working people to appear equally adept. To me that can lead to problems.
Writing and researching , the act of me writing things down helps me learn. I guess it’s kinetic or something. I hate to think of having people in leadership who didn’t really learn it they just did AI
I can barely contain my dismay. The time spent researching and writing a PhD is not "wasted", and nobody who believes that it is has any business at a university. This man's attitudes are not a solution -- they are the problem.
This was surely a discussion without a direction. Neither Boghossian nor Gherman address the foundational questions. One might begin by asking "What is thinking?" "What is understanding?" One might ask the Aristotlean question: "What is the goal or purpose of education?" Here is another question that both Aristotle and John Dewey ask: "What is the relationship between education and political life?" Perhaps the lack of seriousness and depth in this discussion is itself symptomatic of what happens in an AI culture. Those truly interested in the relationship between education and becoming a human being should read Martha Nussbaum's "Cultivating Humanity." Finally, Gherman betrays his abysmal ignorance of the history of science and technology when he casually dismisses the mere "writers" who aren't "doers." Space travel wouldn't be possible if it weren't for the mere "writing" of a fellow named Isaac Newton entitled Philosophiae Naturalis Principia Mathematica. By the way, the founding ideas of computer science are found in the mere "writings" and articles of a philosopher and logician by the name of Alan Turing. Maybe Gherman skipped those classes.
It was bound to happen.This is not going to go away. And all of my criticisms of this tech, and Joe, I had in the first ten minutes, they hit on by the end of the convo. After 37 years in higher ed. I've had the same criticism of higher ed. and honestly, I agree with Joe about the state of higher ed. and the situation students find themselves in in these "systems". Thankfully there are still really good teachers and institutions out there. StealthGPT will throw a HUGE monkey-wrench into it all. It's an "evolutionary" outcome of the culture/tech we live in, even if I don't really like it. One possible saving grace I though may be useful, or not, is when humans write their writing often has a "voice", that voice AI may not be able to mimic. If the reader is familiar with the voice of the supposed writer, there's the detection. And don't think that most writers will be able, or want to, work that hard at "prompt craft" to teach the AI how to mimic their "writer-voice". I've said too much.
Here’s the problem I have with the comments section: You know how to drive a car, operate machinery, use a calculator, and handle a smartphone. But 99% of people don't need to understand how any of these gadgets work in depth; they just get the job done. To the point being made, the current education system is like a driving license provider that demands you understand every component of a car before granting a license, even though you already know how to drive. This is completely unnecessary. Many argue we’re overwhelming students with irrelevant information and requiring them to produce extensive papers that they won’t remember in the long run. When you overload bureaucracy with pointless details, it’s only a matter of time before someone finds a way to sidestep that excess. Many comments suggest that the education system is flawless, created by some god-like being. But in reality, overloading the human brain with excessive, irrelevant knowledge leads to no useful outcomes. The current education system was designed by people decades ago and has not seen significant upgrades. The ultimate goal in education is to equip you with problem-solving skills-that's what people are paying for. They don’t care how you achieve those skills as long as the results are satisfactory. If you play video games, you might be aware of the trope in the Dark Souls community that says if you haven't beaten the game with level 1 stats, or used a shield or used online help, then you haven't truly completed it. The point is that you only need to be competent enough to complete the task, not necessarily the best at it. The education system is flawed, and now people are using AI to exploit those flaws, which is causing concern. Inventions often arise from accidents or from copying ideas, yet ultimately, it’s the invention itself that matters. If you insist that a player has truly completed a game only by fulfilling all side objectives, even when the game only requires completing the primary objectives to roll the credits, then you're being unreasonable. History has shown that many inventors had limited knowledge of a subject, yet with that knowledge, they changed the world. That is what truly matters in the end. It's similar to bureaucracy: while we all need it, too much of it can backfire on the very people it was supposed to help. PS-The current education system emphasizes grades and numerical scores over learning and problem-solving skills. Considering the human brain's tendency to seek shortcuts and conserve energy, it's not surprising that students might prioritize achieving high grades over the actual acquisition of knowledge and problem-solving abilities.
Additionally, if you're attacking someone's beliefs, as in the case of the guest here, you just cant say well this is not how orthodox education is supposed to work and label someone a dishonest idiot. Instead, it's important to delve deeply and deconstruct their argument by offering constructive criticism. Otherwise, it comes across as immature, akin to the behavior of a 16-year-old on Twitter user.
I blame the parents. Nah, but seriously, he's probably a psychopath but... who knows? Some good might come of it. Society might collapse in the meantime, but it'll all be worth it...
Visiting faculty at New College (who was there, commenting because the person mentions this): New College has a narrative system [internally, faculty give "satisfactory" or "unsatisfactory"], it isn't "pass/fail" [it functionally is similar, but it isn't 50% pass/fail typically]. Regardless, the school is working on a way to translate their narrative system to a grade-based system as other schools don't use this approach. How professors determine if a student provides satisfactory work is entirely in the hands of the professor, that includes marking work of students [which is what I did]. One major drawback of the narrative evaluation approach is you can't distinguish between students closer to exceptional students versus mediocre students in a way that can appear on their transcript, and has a bar minimum way to do this; this is flawed, as it may encourage favoritism from professors who do not understand their responsibility to academics [which I think is a problem with a subset of the present faculty [some of the legacy faculty] at that school]. This doesn't still address the cheating concerns, I had students still cheating there [in Computer Science] nor does it address incompetent academics. It's just another way to evaluate students. It doesn't address maintaining standards. Submitting to "grade inflation" as an issue [which is a real thing] to ditch grades [which is a tool] is kind of like throwing away a valuable tool you have in your toolbox. Evaluation becomes much more subjective [and potentially less accountable] when you stop measuring performance [as to determine if a student understood the exercise]. Maintaining standards is a demand of a culture, the answer isn't to just throw it out the window but to uphold them and hold people to account. How do you know students learned something? That's what assessment is to help you determine, to determine if they have earned the academic cred.
I dont disagree with some of his intentions but hes going to need to be much clearer on his ethical position if he wants to be taken seriously. Hes not considered the intellectual implications or the point of a lot of the positive qualities of academia which this will also destroy. He's created an entire shortcut to being able to think, and the loss of all the capacities that thinking well brings.
It's moments like this that I am glad to work in AI Ethics. Humanity remains to be the issue here. An AI can "fake" earning a PhD, but it cannot create knowledge beyond what is in its training data. I have debated before that many discoveries are to be made by collating our current knowledge using AIs - that's fine and good - but it should always be apparent where the AI is involved. Faking that the AI's work is your work is simple fraud by misrepresentation and should, frankly, be a police matter.
Exactly. AI can only plagiarise from thoughts generated by humans. I am an artist so from my pov, AI cannot come up with new ideas or art styles, it cannot inject emotion, intention and lived experience into the images it makes, it just scrapes from the already existing work of human artists and mashes it together like a very clever type of collage. AI is excellent at colour theory and composition but It cannot create anything new. If we let AI take over the creative thought process we will stagnate.
@@moniquevanderpol3656 The issue is, essentially, what else do humans do? It's mostly the same. We just have a larger pool to draw from. It will not be long before AI can do intuitive leaps, and then we will be in Bladerunner territory, frankly. The nature of art (a specialist subject for me) is that you must be able to fail. We do not say that a printer creates art. It simply reproduces it. It cannot "fail" artistically. The nature of AIs is making us challenge that idea; it asks - "What if I can look as if I can fail?" is it art, then? When AI augments humans, is it art? Do we remove the art from a writer using a grammar checker? Do we take the art from Hemingway because he used an editor? Art is that thing that perfectly meets form and function. It is the coin's edge, between and subliming either face of expression or intent. I can intend to hit, and I can express my body's power through a style of hitting, but it is the meeting of these two things at the moment of impact that makes it a martial art. Same with painting. My paintings need to be able to fail to be art. I did a whole series of watercolours using a digital stylus. I could undo my mistakes. I could zoom in, do layers and I could dry the "paper". Was it art I created? What if I got an AI to use the same program and to learn from first principles how to master it. From getting it wrong to learning and performing feats of creation. Does this form art? This is a subject I am considering for Phd.
Even 'undetectable' AI is detectable. The AI won't detect it as AI, but any experienced reader will realise that something is wrong with it. AI writes in a very vague manner, it's basically just spewing out words and sentences related to the topic. To make it 'undetectable', it will paraphrase in a weird way and insert imperfect synonyms. It basically reads like a high school student trying to inflate their word count. In a way, I agree with this guy. The fact AI is so successful in the university, is an indictment of the university itself. Most academic papers/writing is already so vague and puffed up, that it is indistinguishable from AI.
It's getting scary how quickly it's getting better. I would say that what you just wrote would be correct two years ago, but not so much today. I read an article the other day, and it took me almost to the end before it dawned on me that it was AI generated. A couple of years ago, you could tell by the end of the first paragraph.
It is sad that Boghossian, a philosopher by training, failed to recognize and to pursue the obvious parallels between stealthgpt and the story of "The Ring of Gyges" in Plato's Republic. People can already cheat with AI, but with Gherman's product one can cheat without detection. Sadly, Boghossian doesn't seem to recognize that Gherman is just the tech bro version of Thrachsymachus. Gherman seems to think it is a good thing for the person who cheats to cheat and get away with it. He is proud to facilitate cheating with his software that allows one to cheat undetected. Why doesn't Boghossian criticize this moral nihilism more strongly? Perhaps because deep down inside Boghossian is also a moral nihilist, despite his faux persona advocating Socratic inquiry.
In video game terms, if your character has spawned with built-in good memorization skills while another character hasn't, this will level the playing field, as everyone's brain works differently.
Impressive technology, thought provoking discussion. If the user just parrots a question and turns in the output, I can't see a positive benefit. If the user is trying to solve a big problem and has to formulate the question, then the answer is a step forward in understanding. But if the user does the research themselves, they will learn a lot more than the content that actually makes it into their paper. It's this tangent knowledge gained that is lost when the research step is skipped, not to mention the value of the research process itself. I wouldn't want my doctor to skip the time-proven process of doing the research.
00:32:00 At the university, I had a course on the history of my stem discipline, and in it we had to prepare a paper reviewing the history of some specific view (theory, hypothesis ...) . I've spent quite a bit of time going into literature dating back to the 19th century to prepare mine. I've found it instructive, and this probably had an impact on my subsequent work. You can't do anything like that in class, and it would be sad to see such things gone...
@@miramichi30 That's true. Although it did SOUND as if Peter was endorsing him. Maybe he was just being polite in his conversation. I do think Jozef is fundamentally wrong. I hope my doctor doesn't get all his essays written on AI. I want him to know what he's doing. Jozef's just making it easier for people to cheat.
Is the essay a good learning/evaluation tool? Is the AI writer a bit like the calculator/spreadsheet which replaced log tables and hand calculation in maths?
Schools should focus on helping students develop the ability to competently discuss and debate ideas, rather than simply memorizing and regurgitating information. But then again, personal ai tutors will serve this purpose by giving students feedback on their responses, verbal or written.
Two things can be true, the university can be corrupt and also this kind of AI cheating is not helpful. How come no one is asking what the actual purpose of the university is supposed to be? it used to be to gain knowledge and to learn. The goal of the university should not to set you up for a career in finance. although that's what it's become. They are simply degree mills and education is absolutely secondary.
@@gtrdaveg Do you USE ai yet? Or do you still misunderstand the phrase "written with" because to me this fear sounds like the papyrus people worrying about public libraries filled with books. This is OFFLINE man. You can shape any KIND of AI and tell it how to behave. If you are unethical you can be unethical. Just like a book. But ... AI literally just outted CURRENT ivy league faculty for rampant years and years of plagiarism. So.... meet the other edge of the sword.
It will be good for academia not because it will expand it, but because it will shrink it. Currently almost 50% of many countries attend university. University isn't for 50% of people and never was. If we're lucky, AI will finally lead to the culling of the cruft within universities that's built up over the last 100 years when universities became employee factories rather than the centres of academic excellence they're supposed to be.
@DylanYoung I agree that that's good - and maybe I'm being a little pedantic, here - but it's good because it's bad for academia. Academia is an industry. Less students means less money, which is bad for the industry, which is good because the industry is corrupt.
Good luck with your viva voce if you had AI write your computational QM chemistry thesis. Otherwise, I don't really have a problem with it. The garbage degrees will become even more worthless and the real degrees will still require students to demonstrate deep knowledge.
The problem is that it pollutes the information space of these disciplines, wading through seas of AI bullshit to find a single piece of original work makes it less productive to participate in. No degree will be safe when it only took AI a few years to catch up to college level math, coding, law, and other technical fields, and it's only continuing in that direction.
Well I would say he just made the crisis far worse BUT…. Taking your notes and lapsing them into a program that will help you study and generate test prep. That is cool the rest sounds like a really bad idea
Point of order: at 3:30, Gherman says, "I know you [Peter] are a Ph.D." No comment from Peter. Are you a Ph.D? Because Wikipedia has you as an Ed.D. And while Wikipedia is notoriously unreliable, I also have a copy of your publicly available dissertation, which says you are an Ed.D. Care to clarify?
I understand the criticism and reservations others may have for the implications of technology like stealthGPT, but I personally agree completely. I think the real solution to reforming/rebuilding academia (as difficult as it may be) is to encourage intrinsic motivation to learn in students - to actually learn to excel, vs earning a grade. I know that might sound naive and obviously the need remains for standardized assessments of competence, but I think this is a solvable problem as well - eg the in-class essays Peter mentioned, proctored exams, or in the case of software development (my own field), in-person technical interviews.
Wow. This has enormous implications for academia. If I were a college professor, which is worse - knowing that the majority of your students are cheating or that what you are teaching is considered worthless by “modern” standards? Gosh, how depressing.
Can I sue AI companies for the balance of my student loans? Since all degrees will be worthless now. Also, all those students, etc. who wrote essays used to train StealthGPT, were they remunerated?
Hes not solving anything, he's just breaking them. The other app that he talked about that helped build tests from notes sounds much better, but hes shilling product without having really thought through the implications
I think the one question Peter didn't ask him is in regards to competency. If society is upheld by competent people, and this product allows people to skip competency tests, as it were, then don't we just end up with a lot of Dunning-Krugers and cheaters? Try going to Africa or Latin America where corruption runs rampant from the common man to the President, thereby creating and perpetuating cheaters in the system, to the point that it's not about what you know but who you know (nepotism, favors, etc). Jozef somewhat touched upon the competency issue when suggesting a new dynamic emerge, and lauding doers over rule-followers. But philosophically-speaking, we should extrapolate and look at it from a perspective of "rule utilitarianism or possibly Kant's categorical imperative" (thanks ChatGPT for reminding me of these terms, lol).
I imagine this will mean that viva voce will become the essential component for everything. If you can’t get up and speak and answer questions about your topic, you can’t display expertise. I suspect that writing a clear argument will be necessary for that. We might all just go back to the agora and debate rather than be stuck on the publish-perish treadmill. But the writing component will be for personal growth not public display. Could be good 🤷🏻♀️
0:31:00 But should a university be oriented towards that majority? Or should it rather teach (preach) the fundamentals hoping that there will be in the audience one student who really needs this detailed understanding to later advance our collective knowledge?
That guy's claim that he can make an undetectable paper generator is misleading. There is no good way to detect AI written content, but there could be a powerful AI detector created in the future.
If someone uses it to get through, attain tenure, and then teaches students who do the same, what's the point of the whole process? This is a sliver of the larger conversation. If AI can sufficiently do the job, there will be far fewer roles for humans in that job market. People opposed to using AI in its infancy are the same as those who were once against calculators and spell check.
"It's a useless piece of toilet paper". Only, what you learn to be able to obtain that paper, is what validates that paper. Thus he's working towards making that paper useless: which is totally fine if that's his goal, he should just be upfront about it. And I say this as someone who isn't likely to earn a PhD (never say never), but I don't view PhDs as I would, say, a Bachelors. I associate a PhD with an intrinsic yearning to learn and grow. Something that I truly feel that he...may lack? And again that's also entirely fine. I'm not here to run a purity test. But it's his dishonesty that irks me.
This could be a good thing for academia. If everyone can just memorize and repeat the answer to score an A it becomes meaningless. Change all the repeating the answer people to a C and make the people who come up with something truly unique get an A. That will mean you have to outsmart the AI generated papers if you want a good grade. It is like a hammer. Anyone can buy a hammer and put a nail in. Those are C grade people. If you want to be an A grade you need to invent a new type of hammer that does the job better.
Perhaps it’s time to change the way to test student knowledge acquisition rather than “writing papers”. Can anyone out there think of a better way? Suggestion welcome. How about, on ‘test day’ a student is sat in a room with an “isolated” computer and see what they can produce in 8 to 10 hours.
Rufo didn't do shit, it was an Austrian by the name of Dr. Stefan Weber ❤❤ he specialises in this and has exposed many European leader including German Anna-Lena Baerbock. She unfortunately seems to have skated by any accountability 😢
Why would you do that?? Despite all the Ai safety/alignment stuff being done by big companies- you got guys like this trying to make a name for themselves and screwing it up for everyone else.
Using that sort of AI is just gonna make people dumb/not think for themselves or be creative and add their personal thoughts/personality/communication style in their writing. Some of my colleagues use AI for writing emails ...and you can spot it a mile off
He claims his product to be able to help you scam you professors in accepting work done not by you but by AI. Why would you believe that he is being honest with you? The problem with AI is alignment. That is that the AI may well be trying to do something different from what you intended it to do. What you want is a good essay that will withstand checking by a professor. Because of the way they are trained AI tend to give answers that are pleasing on initial reading - but the facts will often be made up and the references are often made up. Nothing he said suggests his stealth AI has solved this problem. The worst that can happen with a chatbot is that you get found out or just as bad an essay full of fake references. The alignment can have much more serious consequences if we give AI more important tasks than writing college essays - potentially deadly.
@@NightsideOfParadise there is no consciousness and they can’t even reason yet. They take a bunch of data and make it easier to understand. His apocalyptic scenario is science fiction and he seems to believe it isn’t. We have much more problem to solve that exist. AI could be helpful to solve them.
For sure you could write an Ed.D. for say Claudine Gay, but avoid the plagiarisms. I don't see AI as a problem if the student can also demonstrate the understand the content the AI generated. I have chem students use AI to do homework, but if they can't replicate the problem solution on an exam the are AI-F-ed. Science still requires understanding and math ability away from your devices.
Peter, this is the first video of yours where I have been disappointed with your engagement on the topic. There was too much focus on this being seen as cheating and not enough on the impact it will have on our ability to attain knowledge (i.e. the dumbing down of society). At the end of the day it is a tool that helps people complete a task. I am sure some said when it launched that using the internet to complete assessments was “cheating”. These tools are important for disrupting an outdated system for measuring whether someone will be useful in contributing to a common enterprise, e.g. a workplace or a field of research. Assuming the tools are accurate (which is still up for debate for AI), then integrating them into our social enterprises should help us advance. As your guest said, the question is what do people want to get from university. If it is simply a job, then it is an outdated model, so people should game the system until employers realise as such. If it is to advance knowledge, then we need to be better at evaluating someone’s knowledge (as a prerequisite to them being trusted to contribute at a particular level) than “open book” assessments/papers. Your endeavour to build a different academic system to the current one (which I admire) should not just use the same structure but with integrity. You should really question whether the current approach to academia can be redesigned, so that among other things these tools can be seen as a positive contribution to the pursuit of knowledge, just like other tools as simple as a calculator are today. At the end of the day we need to start developing systems that help us determine how much we can trust information we consume, whether that is from human or non-human sources. There is definitely some great work being done in this field (e.g. through blockchain and cryptography), if you would be interested in exploring it.
What is this guy going to do when lawsuits start coming in for various forms of fraud, plagiarism, dishonesty and misrepresentation? He has designated himself judge/jury as to what an institution decides is of value/required. If you don't value the university's degree, then don't go there. Sooner or later, universities will put a stipulation (if they haven't already) as part of their academic dishonesty codes that if you are found to have used something like this that the university reserves the right to revoke your degree and slam you with the same sort of academic dishonesty penalties you would receive if you're caught cheating.
Criticise the guy all you want. Yes its unfair. Yes its uncomfortable. But AI is here to stay, whether you like it or not. Essentially, you are complaining about people using cars, while you are sitting on your horse.
Terrible analogy....this is far more like a bunch of morbidly obese people paying a taxi driver to take them to the end of the marathon and then claiming they're the healthiest people in the city. You might get away with the scam but you'll still be fat and die early. People taking the shortcut are failing to develop their intellectual skills in a fundamental way. "But university doesn't teach me the things I need to know..." Its not about the specific skills and knowledge, it the process of learning how to take in, organize, synthesize and restructure ANY field of information, and of learning how to manage your own time and set clearly structured long-term goals (among many other things) that make a difference between those who have acquired a university education and those who merely turned up for a few years and got a piece of paper at the end.
14:27 So we have a system of academic phonies and this guys response to that is, let’s create a phone free for all? I mean fine, let’s burn the whole system down but, do we have a parallel system which is functional to go in its place? Otherwise this guys just a troll wanting to watch the world burn.
![ตีสิบเดย์ [FULL] | เปิดใจ "อาจารย์ เบียร์ ฅนตื่นธรรม" นักสนทนาธรรมสายดุดัน!](http://i.ytimg.com/vi/vqaBI_mnoDk/mqdefault.jpg)
This guy has a degree in intellectual dishonesty.
Can you explain
Which guy?
So strange when you encounter people with no moral compass who just have to kinda fake being a real person. Thanks for having him on.
no true Scotsman
@@Ghost-pb4ts sick non sequitur, bro
@@yyguuyg If u say so
This guest has contempt for honest people 🙃
A world without friction or struggle is a terrible place to live in. The obstacle is often the way to a meaningful life.
I'll tell my doctor at my next colonoscopy 😂😅
@@drkzilla funny !!
Reclaiming their time from reading, researching and learning.
Interesting.🤔
So if it is not cheating or dishonest, why have it be undetectable?
It's not cheating because everyone does it, even his parents 🤪
@@yyguuygthe cope is real
Dr. Boghossian,
It’s unclear to me what ‘AI Will End Academia’ means. I’m - presently - unimpressed with AI. Seven examples:
1. My university-level philosophy courses required interaction with professors and students - often employing the Socratic Method. The inability to do research without AI assistance would be apparent.
2. Many of my law classes employed the Socratic Method.
3. In law school many final exams are based upon questions / cases never covered. It’s to see if the student understands how to correlate what’s been learned.
4. In a recent case, two lawyers provided information to the court. The information / references used in their argument(s) were generated from AI. They were disbarred.
5. I’ve been a Commercial Pilot and Flight Instructor for decades. It takes me seconds to formulate open ended questions stumping AI - questions which don’t stump well trained students. Examples: What is the minimum medical certificate required for an FAA certificated flight instructor? Why? (To date, Dr. B., AI cannot answer this simple question.)
6. When I renew my instructor certificates (every two years), I do this in a classroom setting surrounded other instructors and generally taught by a Ph.D. Why? ‘Learning / Teaching is an active process.’
7. What would happen if I asked AI, “What does Dr. Peter Boghossian mean when he says ‘AI Will End Academia’?” Will I get the correct answer - one you’ll sign off on? Maybe better to listen to the video and, on occasion, interact with you? :)
yYM
Thank you. Unlike Boghossian and Gherman who sound like thoughtless Chatbots, your response reflects serious creative thinking.
You are right, and AI has many more faults on top. Yet, as an academic myself who 'plays a lot' with AI, I found pretty good use cases for it. For example, I was writing a European research grant proposal where I was forced to write in extensive newspeak passages about how my research advances EU priorities, etc. I uploaded a 100-page pdf full of EU political newspeak and the science part of the grant and instructed the AI to find matches and rationalize them. AI did brilliant job at that, saved my time, and lot of anger I always get from reading half-witted bureaucratic nonsense.
Just for fun I put that prompt into ChatGPT. If you happen to read this comment Dr. Boghossian, I'd be curious if you'd give your opinion, perhaps on a scale from Strongly Disagree to Agree, on how much ChatGPT's response reflects your real beliefs.
When Dr. Peter Boghossian says "AI will end academia," he refers to a belief that AI will disrupt the current academic system, which he sees as being overrun by ideological conformity and a lack of rigorous critical thinking. Boghossian has been critical of modern universities, which he argues have become dominated by "woke" ideologies that stifle debate and free inquiry. He feels that academia is more focused on promoting certain ideological positions rather than fostering open dialogue and the search for truth.
AI, in his view, could further expose and challenge the deficiencies of current academic practices, especially in areas like peer review, where low-quality or ideologically-driven scholarship sometimes gets through. He’s suggested that the automation and analytical capabilities of AI might replace many traditional academic roles or bypass universities altogether, potentially leading to their decline.
By contrast, Boghossian supports educational environments that prioritize open discourse, critical thinking, and diverse viewpoints, which he sees as lacking in modern universities.
@@mitzee8621 This is incredible. AI is getting better at an alarming rate. Even a couple of years ago it was so easy to spot AI written articles, now it's much more difficult, and soon it will be impossible.
@@miramichi30 It's already well beyond my ability to spot, with the exception of some technical topics I know a lot about. I don't have much of an eye or ear for AI generated images and music either. If I don't see the person actually creating it in front of me I always harbour doubts as to it's authenticity at this point.
The fact that universities are corrupted institutions that do not punish plagiarists, etc., is actually not a reason to jump further down that hole. I don't want to live in a world where every other engineer and doctor has cheated their way through school using AI and is fully unfamiliar with the fundamentals of their field, sorry. That shit might be fine for the English department but some disciplines actually matter for human lives.
The problem is no different than a factory worker. If you can build a machine that can do the job flawlessly and faster than the best human worker then why do we need them? The factory workers need to learn to build or maintain the machines which destroys one job and creates a different new job for people to do.
The only time cheating hurts is when someone can cheat to get a piece of paper that says they can do something when they can't really perform the job. If a computer can do a job better than the human and that computer can be used to replace the real human in the real job then there is no reason for the human to learn to do the job in college that they will never use in the real job. Now this isn't normally a end all learning of one thing because computers can do it now but the requirement to do that part in college needs to be slowly reduced so instead of spending 100 hours learning it they only spend 20 hours learning it so they still have some knowledge of it as a failsafe but knowing they will likely never need to know it in their job but still have enough knowledge to deal with it for the rare case it is needed.
As long as a deep knowledge of a subject is needed, then a way to check that everyone really understands their subject is essential.
Any student who submits a thesis, study or any other test should be asked questions about it to determine if they understand what they submitted.
For example, "This paragraph you wrote, what do think the main point of it was?" How does that tie in with what you wrote here?"
Those sort of questions will quickly determine if they cheated or not.
I think AI is great as a practical tool, but not for a way to determine understanding of a subject. For example, I use chatgpt to run through some ideas for inventions that I have thought about.
@@TheSimCaptain I agree there should be a way to see if they really know the stuff. However, in today's world everyone is sitting there with a smart phone and access to the internet. How about to be more realistic allow them to search the internet for an answer to any question you ask them. Think how much more realistic it would be for smart phones and internet access to be allowed when taking a test.
As a software dev I don't have every single coding language memorized but I know how to search and find the information I need to write the code. I rarely touch C++ but I have a few times used it to write hacks for games I play to get past long boring parts so I can do other things.
I think the ability to look up and find the information you need is the most important skill for almost any job. Not everyone needs to memorize everything but they need to understand the basics and know how to find the information to complete the job.
Um, doctors and engineers have already been taken over by cheaters.
Why do you think we have had the feminist bridges that collapsed?
All the migrants that was claimed to be nurses and doctors, but without basic western understanding of medicine.
All of them bowing to the ideas of DEI. Acknowledging their privilege of their skin color.
Non of these institutions are not already corrupt.
It's mostly an issue of assessment validity. All these tools are producing the type of output that students are assessed on. Change your tools of assessment to those which can't be done by AI (personal interviews, in-class assignments...) and you'll be just fine - your degree will be a reliable indicator of competence.
You can almost feel the narcissism and the sociopathy through the screen... Wow.
The irony is the guy probably wouldn't have been able to build this thing if he already had it becasue he would not have learned how to.
Why do you assume he has built it? Not just plagiarized somebody else work?
Or been given it by one of the 3 letter agencies, who do collect user info, to be used for a later date blackmail.
@@haraldbredsdorff2699 Haha that's such a great point.
That assumes that people only learn things because of academic requirements. There are plenty of people who study and learn things simply because it interests them.
I have learned the most in papers I've researched and classes I've taught. "Memorization game" is the opposite of this - he's undermining his own point. I feel like stealth-ai wrote all his responses. Also, please explain how science degrees (especially) are worthless.
This is not good. This guy is a straight up villain. Engineers don't need more time to fuck around, they need to study and understand things in a deep technical sense, not pass the work off to an AI.
Well said.
So he is against the current system of grading where people strive for an A+ but is okay with his AI product for "learning" and grading in which the student partied all night and plugged in his product and in 10 minutes gets a paper for an A. Oh brother! I can see his point if the student does not value the class and is only going through the motions because it is a requirement for a degree so the system would need to be changed. However, it still allows less talented and/or hard working people to appear equally adept. To me that can lead to problems.
Writing and researching , the act of me writing things down helps me learn. I guess it’s kinetic or something. I hate to think of having people in leadership who didn’t really learn it they just did AI
I can barely contain my dismay. The time spent researching and writing a PhD is not "wasted", and nobody who believes that it is has any business at a university. This man's attitudes are not a solution -- they are the problem.
This was surely a discussion without a direction. Neither Boghossian nor Gherman address the foundational questions. One might begin by asking "What is thinking?" "What is understanding?" One might ask the Aristotlean question: "What is the goal or purpose of education?" Here is another question that both Aristotle and John Dewey ask: "What is the relationship between education and political life?" Perhaps the lack of seriousness and depth in this discussion is itself symptomatic of what happens in an AI culture. Those truly interested in the relationship between education and becoming a human being should read Martha Nussbaum's "Cultivating Humanity." Finally, Gherman betrays his abysmal ignorance of the history of science and technology when he casually dismisses the mere "writers" who aren't "doers." Space travel wouldn't be possible if it weren't for the mere "writing" of a fellow named Isaac Newton entitled Philosophiae Naturalis Principia Mathematica. By the way, the founding ideas of computer science are found in the mere "writings" and articles of a philosopher and logician by the name of Alan Turing. Maybe Gherman skipped those classes.
This is absolutely ridiculous. It keeps all users from actually learning anything.
such a very very bad idea
The slight smile on this guy is really annoying. There's immense value in learning and struggle.
It was bound to happen.This is not going to go away. And all of my criticisms of this tech, and Joe, I had in the first ten minutes, they hit on by the end of the convo. After 37 years in higher ed. I've had the same criticism of higher ed. and honestly, I agree with Joe about the state of higher ed. and the situation students find themselves in in these "systems". Thankfully there are still really good teachers and institutions out there. StealthGPT will throw a HUGE monkey-wrench into it all. It's an "evolutionary" outcome of the culture/tech we live in, even if I don't really like it.
One possible saving grace I though may be useful, or not, is when humans write their writing often has a "voice", that voice AI may not be able to mimic. If the reader is familiar with the voice of the supposed writer, there's the detection. And don't think that most writers will be able, or want to, work that hard at "prompt craft" to teach the AI how to mimic their "writer-voice". I've said too much.
thanks someone here get the point
Claudine Gay so needed this!
Yes!
Here’s the problem I have with the comments section:
You know how to drive a car, operate machinery, use a calculator, and handle a smartphone. But 99% of people don't need to understand how any of these gadgets work in depth; they just get the job done.
To the point being made, the current education system is like a driving license provider that demands you understand every component of a car before granting a license, even though you already know how to drive. This is completely unnecessary.
Many argue we’re overwhelming students with irrelevant information and requiring them to produce extensive papers that they won’t remember in the long run. When you overload bureaucracy with pointless details, it’s only a matter of time before someone finds a way to sidestep that excess.
Many comments suggest that the education system is flawless, created by some god-like being. But in reality, overloading the human brain with excessive, irrelevant knowledge leads to no useful outcomes.
The current education system was designed by people decades ago and has not seen significant upgrades. The ultimate goal in education is to equip you with problem-solving skills-that's what people are paying for. They don’t care how you achieve those skills as long as the results are satisfactory.
If you play video games, you might be aware of the trope in the Dark Souls community that says if you haven't beaten the game with level 1 stats, or used a shield or used online help, then you haven't truly completed it.
The point is that you only need to be competent enough to complete the task, not necessarily the best at it. The education system is flawed, and now people are using AI to exploit those flaws, which is causing concern.
Inventions often arise from accidents or from copying ideas, yet ultimately, it’s the invention itself that matters. If you insist that a player has truly completed a game only by fulfilling all side objectives, even when the game only requires completing the primary objectives to roll the credits, then you're being unreasonable.
History has shown that many inventors had limited knowledge of a subject, yet with that knowledge, they changed the world. That is what truly matters in the end.
It's similar to bureaucracy: while we all need it, too much of it can backfire on the very people it was supposed to help.
PS-The current education system emphasizes grades and numerical scores over learning and problem-solving skills. Considering the human brain's tendency to seek shortcuts and conserve energy, it's not surprising that students might prioritize achieving high grades over the actual acquisition of knowledge and problem-solving abilities.
Additionally, if you're attacking someone's beliefs, as in the case of the guest here, you just cant say well this is not how orthodox education is supposed to work and label someone a dishonest idiot. Instead, it's important to delve deeply and deconstruct their argument by offering constructive criticism. Otherwise, it comes across as immature, akin to the behavior of a 16-year-old on Twitter user.
I blame the parents. Nah, but seriously, he's probably a psychopath but... who knows? Some good might come of it. Society might collapse in the meantime, but it'll all be worth it...
Yes, was thinking the same thing throughout the interview, he reads as a psychopath
Well all hail our robot overlords then I guess :)
Visiting faculty at New College (who was there, commenting because the person mentions this): New College has a narrative system [internally, faculty give "satisfactory" or "unsatisfactory"], it isn't "pass/fail" [it functionally is similar, but it isn't 50% pass/fail typically]. Regardless, the school is working on a way to translate their narrative system to a grade-based system as other schools don't use this approach. How professors determine if a student provides satisfactory work is entirely in the hands of the professor, that includes marking work of students [which is what I did]. One major drawback of the narrative evaluation approach is you can't distinguish between students closer to exceptional students versus mediocre students in a way that can appear on their transcript, and has a bar minimum way to do this; this is flawed, as it may encourage favoritism from professors who do not understand their responsibility to academics [which I think is a problem with a subset of the present faculty [some of the legacy faculty] at that school].
This doesn't still address the cheating concerns, I had students still cheating there [in Computer Science] nor does it address incompetent academics. It's just another way to evaluate students. It doesn't address maintaining standards. Submitting to "grade inflation" as an issue [which is a real thing] to ditch grades [which is a tool] is kind of like throwing away a valuable tool you have in your toolbox. Evaluation becomes much more subjective [and potentially less accountable] when you stop measuring performance [as to determine if a student understood the exercise].
Maintaining standards is a demand of a culture, the answer isn't to just throw it out the window but to uphold them and hold people to account. How do you know students learned something? That's what assessment is to help you determine, to determine if they have earned the academic cred.
"How AI Will End Academia"
Narrator: "It won't"
I dont disagree with some of his intentions but hes going to need to be much clearer on his ethical position if he wants to be taken seriously. Hes not considered the intellectual implications or the point of a lot of the positive qualities of academia which this will also destroy.
He's created an entire shortcut to being able to think, and the loss of all the capacities that thinking well brings.
@@RD-fm2wk oh the only implication he cares about is how much cash it will earn him
@@daves6394 It's not about the money, it's about sending a message.
@@daves6394 that's definitely how it comes across unfortunately
@@jozefgherman I got the message that I can pay you to have your AI do my homework.
@@jozefgherman really what message.
screw education,give me your money
It's moments like this that I am glad to work in AI Ethics. Humanity remains to be the issue here. An AI can "fake" earning a PhD, but it cannot create knowledge beyond what is in its training data. I have debated before that many discoveries are to be made by collating our current knowledge using AIs - that's fine and good - but it should always be apparent where the AI is involved. Faking that the AI's work is your work is simple fraud by misrepresentation and should, frankly, be a police matter.
Exactly. AI can only plagiarise from thoughts generated by humans. I am an artist so from my pov, AI cannot come up with new ideas or art styles, it cannot inject emotion, intention and lived experience into the images it makes, it just scrapes from the already existing work of human artists and mashes it together like a very clever type of collage. AI is excellent at colour theory and composition but It cannot create anything new. If we let AI take over the creative thought process we will stagnate.
@@moniquevanderpol3656 The issue is, essentially, what else do humans do? It's mostly the same. We just have a larger pool to draw from. It will not be long before AI can do intuitive leaps, and then we will be in Bladerunner territory, frankly. The nature of art (a specialist subject for me) is that you must be able to fail. We do not say that a printer creates art. It simply reproduces it. It cannot "fail" artistically. The nature of AIs is making us challenge that idea; it asks - "What if I can look as if I can fail?" is it art, then? When AI augments humans, is it art? Do we remove the art from a writer using a grammar checker? Do we take the art from Hemingway because he used an editor?
Art is that thing that perfectly meets form and function. It is the coin's edge, between and subliming either face of expression or intent. I can intend to hit, and I can express my body's power through a style of hitting, but it is the meeting of these two things at the moment of impact that makes it a martial art. Same with painting. My paintings need to be able to fail to be art. I did a whole series of watercolours using a digital stylus. I could undo my mistakes. I could zoom in, do layers and I could dry the "paper". Was it art I created? What if I got an AI to use the same program and to learn from first principles how to master it. From getting it wrong to learning and performing feats of creation. Does this form art?
This is a subject I am considering for Phd.
Even 'undetectable' AI is detectable.
The AI won't detect it as AI, but any experienced reader will realise that something is wrong with it.
AI writes in a very vague manner, it's basically just spewing out words and sentences related to the topic. To make it 'undetectable', it will paraphrase in a weird way and insert imperfect synonyms. It basically reads like a high school student trying to inflate their word count.
In a way, I agree with this guy. The fact AI is so successful in the university, is an indictment of the university itself. Most academic papers/writing is already so vague and puffed up, that it is indistinguishable from AI.
It's getting scary how quickly it's getting better. I would say that what you just wrote would be correct two years ago, but not so much today. I read an article the other day, and it took me almost to the end before it dawned on me that it was AI generated. A couple of years ago, you could tell by the end of the first paragraph.
It is sad that Boghossian, a philosopher by training, failed to recognize and to pursue the obvious parallels between stealthgpt and the story of "The Ring of Gyges" in Plato's Republic. People can already cheat with AI, but with Gherman's product one can cheat without detection. Sadly, Boghossian doesn't seem to recognize that Gherman is just the tech bro version of Thrachsymachus. Gherman seems to think it is a good thing for the person who cheats to cheat and get away with it. He is proud to facilitate cheating with his software that allows one to cheat undetected. Why doesn't Boghossian criticize this moral nihilism more strongly? Perhaps because deep down inside Boghossian is also a moral nihilist, despite his faux persona advocating Socratic inquiry.
Spot on
In video game terms, if your character has spawned with built-in good memorization skills while another character hasn't, this will level the playing field, as everyone's brain works differently.
Impressive technology, thought provoking discussion. If the user just parrots a question and turns in the output, I can't see a positive benefit. If the user is trying to solve a big problem and has to formulate the question, then the answer is a step forward in understanding. But if the user does the research themselves, they will learn a lot more than the content that actually makes it into their paper. It's this tangent knowledge gained that is lost when the research step is skipped, not to mention the value of the research process itself. I wouldn't want my doctor to skip the time-proven process of doing the research.
00:32:00 At the university, I had a course on the history of my stem discipline, and in it we had to prepare a paper reviewing the history of some specific view (theory, hypothesis ...) . I've spent quite a bit of time going into literature dating back to the 19th century to prepare mine. I've found it instructive, and this probably had an impact on my subsequent work. You can't do anything like that in class, and it would be sad to see such things gone...
I like Peter but he was wrong to get this man on. Jozef is enabling dishonesty, lying and cheating.
Having him on is not the same as endorsing him. This is an important development in academia.
@@miramichi30 That's true. Although it did SOUND as if Peter was endorsing him. Maybe he was just being polite in his conversation. I do think Jozef is fundamentally wrong. I hope my doctor doesn't get all his essays written on AI. I want him to know what he's doing. Jozef's just making it easier for people to cheat.
Is the essay a good learning/evaluation tool? Is the AI writer a bit like the calculator/spreadsheet which replaced log tables and hand calculation in maths?
Schools should focus on helping students develop the ability to competently discuss and debate ideas, rather than simply memorizing and regurgitating information. But then again, personal ai tutors will serve this purpose by giving students feedback on their responses, verbal or written.
Two things can be true, the university can be corrupt and also this kind of AI cheating is not helpful. How come no one is asking what the actual purpose of the university is supposed to be? it used to be to gain knowledge and to learn. The goal of the university should not to set you up for a career in finance. although that's what it's become. They are simply degree mills and education is absolutely secondary.
1000% but only as we know it. It wont "end" academia. It will expand it exponentially.
I'm not sure that you're using the term academia in a way that I recognise. I really don't know how Stealth GPT could possibly be good for academia
@@gtrdaveg Do you USE ai yet? Or do you still misunderstand the phrase "written with" because to me this fear sounds like the papyrus people worrying about public libraries filled with books.
This is OFFLINE man. You can shape any KIND of AI and tell it how to behave. If you are unethical you can be unethical. Just like a book. But ...
AI literally just outted CURRENT ivy league faculty for rampant years and years of plagiarism. So.... meet the other edge of the sword.
It will be good for academia not because it will expand it, but because it will shrink it. Currently almost 50% of many countries attend university. University isn't for 50% of people and never was. If we're lucky, AI will finally lead to the culling of the cruft within universities that's built up over the last 100 years when universities became employee factories rather than the centres of academic excellence they're supposed to be.
@DylanYoung I agree that that's good - and maybe I'm being a little pedantic, here - but it's good because it's bad for academia. Academia is an industry. Less students means less money, which is bad for the industry, which is good because the industry is corrupt.
So.....is this not simply a way of outsourcing one's thinking? To what end??????
Teachers can also give you an oral exam to ensure you've understood the subject, which is more about discipline than morality.
Great way to make the in-class essay a staple of classrooms.
Good luck with your viva voce if you had AI write your computational QM chemistry thesis. Otherwise, I don't really have a problem with it. The garbage degrees will become even more worthless and the real degrees will still require students to demonstrate deep knowledge.
The problem is that it pollutes the information space of these disciplines, wading through seas of AI bullshit to find a single piece of original work makes it less productive to participate in. No degree will be safe when it only took AI a few years to catch up to college level math, coding, law, and other technical fields, and it's only continuing in that direction.
Well I would say he just made the crisis far worse BUT…. Taking your notes and lapsing them into a program that will help you study and generate test prep. That is cool the rest sounds like a really bad idea
Point of order: at 3:30, Gherman says, "I know you [Peter] are a Ph.D." No comment from Peter. Are you a Ph.D? Because Wikipedia has you as an Ed.D. And while Wikipedia is notoriously unreliable, I also have a copy of your publicly available dissertation, which says you are an Ed.D. Care to clarify?
I understand the criticism and reservations others may have for the implications of technology like stealthGPT, but I personally agree completely. I think the real solution to reforming/rebuilding academia (as difficult as it may be) is to encourage intrinsic motivation to learn in students - to actually learn to excel, vs earning a grade. I know that might sound naive and obviously the need remains for standardized assessments of competence, but I think this is a solvable problem as well - eg the in-class essays Peter mentioned, proctored exams, or in the case of software development (my own field), in-person technical interviews.
Wow. This has enormous implications for academia. If I were a college professor, which is worse - knowing that the majority of your students are cheating or that what you are teaching is considered worthless by “modern” standards? Gosh, how depressing.
useless things in your life, what like learning stuff...
If students are not allowed to use AI to write papers, then teachers should not be allowed to use AI to set or grade papers - and yet they do.
Can I sue AI companies for the balance of my student loans? Since all degrees will be worthless now.
Also, all those students, etc. who wrote essays used to train StealthGPT, were they remunerated?
I wonder if graders are using AI to read and grade the essays?
We’re gonna see a lot of Elizabeth Holme’s of AI.. this guy is one of them. Just listen to his sentiments and look at his thousand yard stare.
Hes not solving anything, he's just breaking them. The other app that he talked about that helped build tests from notes sounds much better, but hes shilling product without having really thought through the implications
I think the one question Peter didn't ask him is in regards to competency. If society is upheld by competent people, and this product allows people to skip competency tests, as it were, then don't we just end up with a lot of Dunning-Krugers and cheaters? Try going to Africa or Latin America where corruption runs rampant from the common man to the President, thereby creating and perpetuating cheaters in the system, to the point that it's not about what you know but who you know (nepotism, favors, etc).
Jozef somewhat touched upon the competency issue when suggesting a new dynamic emerge, and lauding doers over rule-followers. But philosophically-speaking, we should extrapolate and look at it from a perspective of "rule utilitarianism or possibly Kant's categorical imperative" (thanks ChatGPT for reminding me of these terms, lol).
I imagine this will mean that viva voce will become the essential component for everything. If you can’t get up and speak and answer questions about your topic, you can’t display expertise. I suspect that writing a clear argument will be necessary for that. We might all just go back to the agora and debate rather than be stuck on the publish-perish treadmill. But the writing component will be for personal growth not public display. Could be good 🤷🏻♀️
This was invaluable, what would you do if tomorrow Chatgpt came out with a new "undetectable" model, standing against it serves no purpose
0:31:00 But should a university be oriented towards that majority? Or should it rather teach (preach) the fundamentals hoping that there will be in the audience one student who really needs this detailed understanding to later advance our collective knowledge?
In addition, AI is just basically plagiarism pretending to be original.
Helping students with their studies, nonsense. Like the BS chegg pushed.
That guy's claim that he can make an undetectable paper generator is misleading. There is no good way to detect AI written content, but there could be a powerful AI detector created in the future.
If someone uses it to get through, attain tenure, and then teaches students who do the same, what's the point of the whole process? This is a sliver of the larger conversation. If AI can sufficiently do the job, there will be far fewer roles for humans in that job market. People opposed to using AI in its infancy are the same as those who were once against calculators and spell check.
"It's a useless piece of toilet paper". Only, what you learn to be able to obtain that paper, is what validates that paper. Thus he's working towards making that paper useless: which is totally fine if that's his goal, he should just be upfront about it.
And I say this as someone who isn't likely to earn a PhD (never say never), but I don't view PhDs as I would, say, a Bachelors. I associate a PhD with an intrinsic yearning to learn and grow. Something that I truly feel that he...may lack? And again that's also entirely fine. I'm not here to run a purity test. But it's his dishonesty that irks me.
Where is the struggle? Is the struggle in learning replaced by another struggle?
Would you want a surgeon who produced their own work or one that relied on an AI to answer all their questions?
Now do this for upper level math and see how it does
Who gave Peter the Quaaludes?
This could be a good thing for academia. If everyone can just memorize and repeat the answer to score an A it becomes meaningless. Change all the repeating the answer people to a C and make the people who come up with something truly unique get an A. That will mean you have to outsmart the AI generated papers if you want a good grade.
It is like a hammer. Anyone can buy a hammer and put a nail in. Those are C grade people. If you want to be an A grade you need to invent a new type of hammer that does the job better.
Why assign a paper then? This learning is done through the writing process.
I imagine thatSpaceX engineers do right papers on their work.
00:56:00 I agree with Peter, this feels like cheating, and that's why I wouldn't really want to teach...
Perhaps it’s time to change the way to test student knowledge acquisition rather than “writing papers”. Can anyone out there think of a better way? Suggestion welcome. How about, on ‘test day’ a student is sat in a room with an “isolated” computer and see what they can produce in 8 to 10 hours.
Rufo didn't do shit, it was an Austrian by the name of Dr. Stefan Weber ❤❤ he specialises in this and has exposed many European leader including German Anna-Lena Baerbock. She unfortunately seems to have skated by any accountability 😢
Damn I wish I was a student now.
Why would you do that?? Despite all the Ai safety/alignment stuff being done by big companies- you got guys like this trying to make a name for themselves and screwing it up for everyone else.
Using that sort of AI is just gonna make people dumb/not think for themselves or be creative and add their personal thoughts/personality/communication style in their writing. Some of my colleagues use AI for writing emails ...and you can spot it a mile off
Peter is not me. Woops. 😅
He claims his product to be able to help you scam you professors in accepting work done not by you but by AI. Why would you believe that he is being honest with you? The problem with AI is alignment. That is that the AI may well be trying to do something different from what you intended it to do. What you want is a good essay that will withstand checking by a professor. Because of the way they are trained AI tend to give answers that are pleasing on initial reading - but the facts will often be made up and the references are often made up. Nothing he said suggests his stealth AI has solved this problem. The worst that can happen with a chatbot is that you get found out or just as bad an essay full of fake references. The alignment can have much more serious consequences if we give AI more important tasks than writing college essays - potentially deadly.
If you create a technology that can do the academic work for you, why would anyone hire you? They would just use that AI.
bingo
Apocalyptic predictions can’t be backed up by logic and rationality. Thumb down,
It's not logical to compare pocket calculators and a new form of consciousness superior to homo sapiens.
@@NightsideOfParadise there is no consciousness and they can’t even reason yet. They take a bunch of data and make it easier to understand.
His apocalyptic scenario is science fiction and he seems to believe it isn’t.
We have much more problem to solve that exist. AI could be helpful to solve them.
54:12 This says it all
For sure you could write an Ed.D. for say Claudine Gay, but avoid the plagiarisms. I don't see AI as a problem if the student can also demonstrate the understand the content the AI generated. I have chem students use AI to do homework, but if they can't replicate the problem solution on an exam the are AI-F-ed. Science still requires understanding and math ability away from your devices.
"These papers are toilet paper" 😂😂😂😂😂❤
Peter, this is the first video of yours where I have been disappointed with your engagement on the topic. There was too much focus on this being seen as cheating and not enough on the impact it will have on our ability to attain knowledge (i.e. the dumbing down of society).
At the end of the day it is a tool that helps people complete a task. I am sure some said when it launched that using the internet to complete assessments was “cheating”.
These tools are important for disrupting an outdated system for measuring whether someone will be useful in contributing to a common enterprise, e.g. a workplace or a field of research.
Assuming the tools are accurate (which is still up for debate for AI), then integrating them into our social enterprises should help us advance.
As your guest said, the question is what do people want to get from university. If it is simply a job, then it is an outdated model, so people should game the system until employers realise as such. If it is to advance knowledge, then we need to be better at evaluating someone’s knowledge (as a prerequisite to them being trusted to contribute at a particular level) than “open book” assessments/papers.
Your endeavour to build a different academic system to the current one (which I admire) should not just use the same structure but with integrity. You should really question whether the current approach to academia can be redesigned, so that among other things these tools can be seen as a positive contribution to the pursuit of knowledge, just like other tools as simple as a calculator are today.
At the end of the day we need to start developing systems that help us determine how much we can trust information we consume, whether that is from human or non-human sources. There is definitely some great work being done in this field (e.g. through blockchain and cryptography), if you would be interested in exploring it.
I love it
What is this guy going to do when lawsuits start coming in for various forms of fraud, plagiarism, dishonesty and misrepresentation? He has designated himself judge/jury as to what an institution decides is of value/required. If you don't value the university's degree, then don't go there.
Sooner or later, universities will put a stipulation (if they haven't already) as part of their academic dishonesty codes that if you are found to have used something like this that the university reserves the right to revoke your degree and slam you with the same sort of academic dishonesty penalties you would receive if you're caught cheating.
Criticise the guy all you want.
Yes its unfair. Yes its uncomfortable. But AI is here to stay, whether you like it or not.
Essentially, you are complaining about people using cars, while you are sitting on your horse.
preach
Terrible analogy....this is far more like a bunch of morbidly obese people paying a taxi driver to take them to the end of the marathon and then claiming they're the healthiest people in the city. You might get away with the scam but you'll still be fat and die early. People taking the shortcut are failing to develop their intellectual skills in a fundamental way. "But university doesn't teach me the things I need to know..." Its not about the specific skills and knowledge, it the process of learning how to take in, organize, synthesize and restructure ANY field of information, and of learning how to manage your own time and set clearly structured long-term goals (among many other things) that make a difference between those who have acquired a university education and those who merely turned up for a few years and got a piece of paper at the end.
Faculty discount 😂
14:27 So we have a system of academic phonies and this guys response to that is, let’s create a phone free for all?
I mean fine, let’s burn the whole system down but, do we have a parallel system which is functional to go in its place?
Otherwise this guys just a troll wanting to watch the world burn.
If your house is on fire, do you stop to discuss whether you have a new place to live before calling the fire department?
This guy is a total slimeball. He will either end up a billionaire or in jail.
A calculator is to math as
a LLM is to language.
❤️🍓☺️
I can't believe this guy didnt understand the varied uses of this technology...very disingenuous...
... Because their family is impoverished - Or something like that ... HAHAHAHAHA! This man is a POS ... No way around it.
Just like the calculators would transform kids in ignorants in the 80s? Yeah right!