8 AI companies including OpenAI, Google set to handover their chatbots to hackers. Here's why

livemint.com:

AI chatbots are also vulnerable to hacking. The hackers can misuse the generative AI technology to generate false information, biased information, induce fake narratives, pedal offensive information, etc. Responding to all such threats, the AI companies are participating in 20 hours at the DEF CON conference in Las Vegas, news platform SEMAFOR reported.

The report mentioned that the conference will begin on Friday during which around 3,200 hackers will attempt to explore the vulnerabilities of the chatbots. The Major AI companies like Anthropic, Cohere, Google, Hugging Face, Meta, Nvidia, OpenAI, and Stability AI are expected to participate in the conference and will hand over their chatbots to these hackers. The participation in the DEF CON conference shows that the AI companies are serious about fulfilling the commitments secured by the White House on external testing, the SEMAFOR report added.

During the conference, points will be awarded to achieve various tasks like making the chatbot generate political misinformation. The hackers will also test subtle biases with the chatbot like race or income levels. The development is significant since it can help AI companies to build more secure models which will strengthen their case when it comes to regulation of the artificial intelligence technologies.