Biden’s AI Executive Order

On October 30, 2023, Biden signed an executive order on artificial Intelligence (or AI). This executive order sets forth new standards to regulate AI and directs federal agencies to assess risks and set policies on AI technology. Among many objectives, objectives relevant to startups are: (1) AI Safety and Security, (2) AI Innovation, (3) Consumer Protection, and (4) Consumer Privacy.

AI Safety and Security

The executive order orders the Secretaries of Commerce, Energy, Homeland Security, and other relevant federal agencies to establish guidelines and best practices to promote consensus AI standards and establish appropriate testing procedures, particularly for generative AI and large language models (LLMs).

Further, companies developing generative AI and LLMs must report and record efforts regarding training, developing, and producing AI models, including associated cybersecurity measures and their test results.

AI Innovation

The executive order streamlines visa petitions and applications to AI-related workers and researchers to attract top talent to the US.

Consumer Protection

The executive order aims to protect the public by ordering various federal agencies to establish an AI task force to develop policies on responsible deployment and use of AI in health, transportation, housing, etc.

Consumer Privacy

The executive order aims to mitigate privacy risks associated with AI by tasking federal agencies to evaluate and identify companies that broker commercially available information containing personally identifiable data, and evaluate standards and procedures associated with collection, processing, maintenance, use, sharing, dissemination, and disposition of commercially available information.

Key Takeaway

As with any executive order, the language in the executive order on AI lacks details and precision needed to implement regulations. In the next coming months, affected federal agencies will promulgate draft regulations for public feedback. However, the executive order on AI does offer some insight into how the federal government intends to regulate AI. For example, cybersecurity and protecting AI models against intrusions and hackers appear to be high on the federal government’s concerns. Further, the federal government appears to want to establish guidelines and best practices for AI industry to self-regulate.

If your startup is working on AI-related innovation, now may be a good time to review your cybersecurity policy. And if you don’t currently have one, now may be a good time to draft one. Further, to stay ahead of upcoming regulations, now may be a good time to think about and start drafting a test plan to test for biases in AI models.

Tags: #ai, #regulation