The Biden administration has determined that generative AI models should be evaluated by independent experts. This important task was assigned to the AI Village hacker community who, during the DEF CON 31 conference in Las Vegas (August 10-13), will examine the technologies of various companies to find bugs and other artificial intelligence flaws.
Thousands of hackers challenge AI
Ai Village, founded by Sven Cattell, is a community of hackers and scientists who study artificial intelligence to identify uses and abuses in terms of security and privacy. After receiving input from the White House, researchers will analyze models of Anthropic, Google, Hugging Face, NVIDIA, OpenAI and Stability AI, leveraging a platform developed by Scale AI.
The press release highlights the advantages of the so-called LLM (Large Langugae Model), but also the risks arising from technology (hallucinations, bias, jailbreak) that can become a threat to security and privacy. During the DEF CON in August, thousands of hackers will study the models of these companies through a time-based access. A Capture The Flag (CTF) score will be awarded and those who achieve the highest score will receive a high-end NVIDIA video card as a reward.
Among the participants of the DEF CON conference there will also be hundreds of students. The goal is to understand how AI models work and to improve technology skills. The DEF CON 29 of 2021 hosted the first contest related to machine learning. Hackers examined Twitter’s cropping algorithm to locate the bug that preferred images of women and people with fair skin.
Stripe IPOs now available: open your free account on .
Source: AI VillageInizio modulo
Fine modulo