From Sabine's email for the day: Researchers from OpenAI have put forward an industrial policy for
openai
| Source: Mastodon | Original article
OpenAI researchers have unveiled a draft industrial policy that enshrines a legally recognised “Right to AI,” calling for universal public access to the most capable generative models. The proposal, circulated in a briefing shared by physicist‑blogger Sabine Hossenfelder, argues that governments should fund large‑scale compute clusters and make them available to academia, small enterprises and civil society, thereby preventing a monopoly of power in the hands of a few tech giants.
The move marks a rare foray by a leading AI lab into formal policy design, shifting the conversation from voluntary safety guidelines to a statutory framework. By positioning AI access as a public utility, OpenAI hopes to democratise innovation, reduce the risk of a “AI divide,” and create a regulated environment where safety testing can be performed on parity‑level hardware. The draft also outlines mechanisms for transparent licensing, audit trails and a public oversight board, echoing the European Union’s AI Act but with a stronger emphasis on compute as a shared resource.
Why it matters is twofold. First, it challenges the prevailing market‑driven model that ties cutting‑edge models to proprietary cloud services, a model that has drawn criticism amid concerns over concentration of talent and data. Second, it could reshape funding flows: the policy calls for state‑backed compute budgets comparable to national supercomputing programmes, a notion that may influence ongoing discussions about the $40 billion loan consortium that recently pledged financing to OpenAI.
What to watch next are the reactions from policymakers in the EU, the United States and Nordic governments, where AI strategy is already a priority. If the draft gains traction, legislative drafts may appear in upcoming AI strategy white papers, and OpenAI could pilot a government‑funded compute hub later this year. The proposal also raises questions about how the “Right to AI” will be balanced against national security and intellectual‑property concerns, setting the stage for a heated policy debate in the months ahead.
Sources
Back to AIPULSEN