On 30 October, US President Joe Biden issued an Executive Order (EO) containing an array of directives centred on Artificial Intelligence (AI), ranging from safeguards against algorithmic bias to promoting American innovation in the field. The EO was issued amid a flurry of activity around AI this past year as its widespread integration into software, connected devices, and online platforms has increasingly raised concerns over the potential harms such tools pose alongside their benefits. It also comes as the EU finalises its comprehensive AI Act, which is expected to be adopted at the beginning of 2024, and ahead of anticipated international agreements on AI such as the G7’s Hiroshima Process and ASEAN’s Guide to AI Ethics and Governance.
While the bulk of the EO tasks various federal agencies to develop and publish best practice guidelines on the responsible use of AI in certain contexts, it also provides a clearer picture of the Biden Administration’s priorities in the coming year concerning the technology. One of the more substantial policy directives is a requirement for companies developing foundation models, powerful machine learning models such as OpenAI’s GPT-4, to notify the federal government at the model’s training phase and share the results of certain safety test as well. The EO specifies that the reporting requirement applies to foundation models that are deemed a threat to national security, economic security, or public health and safety. Also of note, the EO explicitly encourages the Federal Trade Commission to “exercise its authorities” as related to the promotion of a fair, open, and competitive AI ecosystem. Taken together, these two directives show that the Biden Administration will be keeping a close eye on the biggest players in the AI field.
Access Partnership closely monitors AI policy developments in the US from ethics to governance and safety. For more information on this topic, please contact Jacob D. Hafey at firstname.lastname@example.org.