Vice President Joe Biden has called on hackers to help keep AI chatbots in check. The White House will back an event at the Defcon security conference this summer that calls on experts to reveal flaws in generative AI systems, along with committing $140 million towards seven new research institutes that will focus on creating ethical, transformative AI for social good. The initiative signals the Biden administration’s attempt to tackle the opportunities and risks raised by AI.
The call for a mass hacking exercise came around the same time as the launch of ChatGPT, a new AI system created by Anthropic and OpenAI. Fear of misuse of AI has risen alongside the newfound faith in its potential to solve problems. Accordingly, European Union lawmakers are currently negotiating a new AI Act that will place numerous restrictions on AI development, while Brazil and China have also proposed forms of regulation to protect human rights.
In the United States, Senator Michael Bennett introduced a bill to set up an AI task force to protect citizens’ privacy and civil rights. Additionally, a group of US regulatory agencies including the federal government promised to use existing laws to ensure the AI is developed properly. Senator Ron Wyden proposed a new law, the Algorithmic Accountability Act, which would require companies to check their algorithms and verify if an automated system is being used.
Arati Prabhakar, director of the White House Office of Science and Technology Policy, said at a meeting in March that AI must be safeguarded if it is going to be beneficial. The Defcon hacking event will aim to carry out this objective, bringing together hackers, policy experts, and companies such as Google, NVIDIA, and Stability AI to align their AI systems to the Biden administration’s AI Bill of Rights and NIST framework.
The AI Vulnerability Database, Humane Intelligence, and Microsoft are also collaborating to lead the hacking exercise. The event will be conducted in the ‘Capture the Flag’ format and it will reward participants for finding and reporting bugs or unwanted behaviour from the AI systems. This hackathon signals the Biden administration’s efforts to bring greater accountability for AI development and secure the technology for public good.
