The MLephant in the Room: Malware Detection in ML & LLM Models

As the demand for AI capabilities grows, LLMs and other ML models are increasingly included in the software that we develop and consume.
Of particular note are Pickle files, which represent 83.5% of all ML models. Their popularity has attracted the eye of threat actors and has become a new attack vector of choice.
To help businesses stay ahead of the threat, ReversingLabs has developed the ability to detect malicious machine learning models and identify unsafe function calls during deserialization, flag unusual behaviors, and automatically classify models exhibiting these behaviors as suspicious or malicious.
Watch this webinar to learn:
✓ What exactly ML models are and why they’re important
✓ The distinction between API and ML models and the role of each
✓ Why ML models are so frequently shared and, unfortunately, abused
✓ The serialization and deserialization process and how it’s used to spread malware
✓ How ReversingLabs has developed the capability to detect ML malware before it can cause harm
About RL:
ReversingLabs is the trusted name in file and software security, to verify and deliver safe binaries. With the largest Threat Repository in the industry with over 40 billion searchable files, the Fortune 500 trusts their software supply chain security and malware analysis with ReversingLabs. Learn more: www.reversingl...
RL - Trust Delivered.
Be sure to subscribe to RL and follow us on social media →
/ reversinglabs
/ reversinglabs
/ reversinglabs
/ reversinglabs