The White House publishes a Blueprint for an AI Bill of Rights

The White House Office of Science and Technology Policy published the Blueprint for an AI Bill of Rights, a document setting out five principles intended to guide the design and use of automated systems: safe and effective systems, protection from algorithmic discrimination, data privacy, notice and explanation when an automated system is used, and the ability to opt for human consideration and a human fallback instead. It was produced under Alondra Nelson, then acting director of OSTP, following roughly a year of consultation with industry, academics, civil-society groups and members of the public.

The Blueprint was explicitly non-binding. It created no new legal obligations, established no enforcement mechanism, and applied only as guidance to federal agencies and, aspirationally, to the private sector — a status the accompanying materials acknowledged, alongside a companion technical handbook of concrete steps for putting the principles into practice. Set against the EU’s AI Act, then moving through the European Parliament as binding law with financial penalties attached, it represented a markedly softer starting position for US policy on the same underlying concerns.

Reaction split along familiar lines. Civil-rights and consumer groups welcomed the principles while pressing for legislation to give them force; industry groups broadly welcomed a document that asked for voluntary adoption rather than compliance. Congress did not turn the Blueprint into statute.

Its principles were carried forward, in substance rather than by direct legal effect, into Biden’s executive order on AI a year later, which added reporting requirements the Blueprint had lacked. That order, and the OSTP guidance underneath it, was itself revoked within days of the Trump administration taking office in January 2025 — leaving the Blueprint’s five principles as an artefact of a particular administration’s approach rather than a durable feature of US AI policy.