The U.S. Food and Drug Administration issued a discussion paper on August 18, 2026 proposing how to regulate medical devices powered by generative AI. It outlines a two-axis framework for risk assessment, a premarket evaluation built on a concept it calls competency assessment, and risk-proportionate postmarket monitoring. The novelty is the reference point. A generative device would face non-clinical benchmarking and clinical confirmation before reaching patients, a sequence modeled on how physicians are trained and evaluated. The agency invites device makers, clinicians, and the public to respond under docket FDA-2026-N-7874 by October 19, 2026.

The medical exam has become an object of regulation. For years a machine read scans and passed suggestions to a doctor who signed the report. Now the agency proposes grading the software itself as if it were a resident sitting for boards. A competency assessment treats the device as a professional subject, with benchmarks standing in for years of training, and that substitution carries a specific risk. Benchmarks measure what the test covers, a fixed set of tasks, a curated distribution of cases, and little about the patient who does not resemble the training set or the decision made in a hallway at three in the morning. The machine can pass and still fail where a physician's judgment would have hesitated. The docket must decide where that gap is measured, and what happens when the certificate holder is software that no one examined in a body.

OpenAI has awarded grants to 14 organizations that will study how public policy should respond to advanced artificial intelligence. The company will distribute $1 million in cash and up to $1 million in API credits across six-month projects in the United States, European Union, Brazil, Singapore and South Korea. The selected work covers worker benefits, taxation, productive ownership, data-center energy, clinical infrastructure, government simulations, legislative oversight, biosecurity and uncontrolled self-improving systems. OpenAI chose the recipients from over 400 proposals, and results are expected in 2027. The organizations are described as independent, while their money and computing access come from a company whose products and expansion some of them may evaluate.

Funding research also selects which questions receive staff, meetings, datasets, prototypes and a place on policy calendars. OpenAI says democratic institutions should shape the response to AI, yet this program gives the company an earlier role in deciding which approaches enter that debate with resources attached. The roster includes the Tax Foundation, the American Enterprise Institute, a Brazilian public hospital and groups working on nuclear and biological risk. Sponsorship does not invalidate their work. It creates a relationship that every final report must make legible. Readers will need the cash award, API credits, model access, contractual limits and any review rights OpenAI retained before publication. Those disclosures should appear beside the findings rather than remain inside the grant agreement.

Spotify will begin marking artist profiles that represent AI-generated identities with an AI Persona badge in mid-September. The label will appear on profile banners and About sections, in search results and beside tracks in playlists. Artists can disclose their synthetic identity through Spotify for Artists, but the company will also review names and profile images and apply the badge itself. A labeled artist can appeal that decision. The badge concerns the public identity presented by the profile. Spotify's separate AI Credits field covers how generative tools were used to make the music. By default, AI Persona profiles will be excluded from editorial and algorithmic recommendations. Their tracks can reach a listener who follows the profile or deliberately searches for it.

The policy creates a class of performer whose recordings can remain available while ordinary circulation is restricted because the face and biography do not belong to a person. That distinction responds to projects such as The Velvet Sundown, an AI band presented as human that accumulated millions of streams. It also gives Spotify authority to decide whether a photorealistic portrait, name and public history amount to an authentic artist. A label may protect listeners from impersonation, yet the decisive action occurs inside the recommendation system. The profile stays searchable while disappearing from Discover Weekly, editorial playlists and other automated routes unless a listener has already chosen to follow it.