President Joe Biden has issued an executive order setting guidelines for generative AI, anticipating forthcoming legislative action. The order outlines eight objectives, including establishing new safety and security standards, safeguarding privacy, promoting equity, and ensuring responsible government use. Various government bodies have been assigned tasks such as developing protective measures against the misuse of AI for biological manipulation and enhancing cybersecurity protocols.
Additionally, the National Institute of Standards and Safety (NIST) will evaluate AI models prior to public release, while the Departments of Energy and Homeland Security are directed to address potential threats to critical infrastructure. Notably, developers of major AI models, such as GPT by OpenAI and Llama 2 by Meta, must share safety test outcomes. A senior administration official clarified that the guidelines primarily concern future AI models and won’t affect existing publicly available ones.
In a bid to protect user privacy, the executive order calls on Congress to enact data privacy laws and supports the development of privacy-preserving technologies. It also aims to prevent AI-driven discrimination and mandates guidelines for various sectors to prevent biased AI use. Additionally, the order instructs agencies to assess the impact of AI on employment and promotes the involvement of more individuals in the AI sector through a National AI Research Resource and expedited hiring of AI professionals within the government.
Previously, the Biden administration introduced an AI Bill of Rights and subsequently forged agreements with prominent AI entities, including Meta, Google, OpenAI, Nvidia, and Adobe. While the executive order is a positive step toward regulating generative AI, it is temporary and subject to change with each administration. Nonetheless, industry experts view the order as a promising development that emphasizes the government’s commitment to AI advancement and regulation.