2024-12-14 –, Track 3
A discussion of the OWASP ML Top 10 and OWASP LLM Top 10, and how a failure to apply these principles in 2001 A Space Odyssey, led to implementation flaws in HAL 9000, resulting in disastrous consequences for the crew.
The talk will use the OWASP Top 10 for ML and OWASP Top 10 for LLMs to anyalze the nature of the flaws in HAL 9000, the AI in 2001: A Space Odyssey, and how this led to disastrous results for the mission.
There will be a discussion of failures to consider different aspects of both the LLM and ML top 10 during HAL's design and training phases, and the subsequent attempts to implement fixes during the mission. Each omission or failure to apply an OWASP principle, that led to the vulnerabilities will be discussed in detail, and also related to real life applications, to ensure the talk isn't just a geeky discussion of a cool-looking scf-fi AI.
Former software developer who became a security consultant. Interested in writing security tools and retro tech.