Two governance moves landed this week: New York now requires ads with AI-generated people to disclose them as synthetic performers, while Anthropic says it has taken its latest models offline to comply with new export controls.
Ads in New York must now label AI-generated 'synthetic performers'
New York has implemented a law requiring advertisements that feature AI-generated people to clearly label those figures as “synthetic performers.” The measure is framed around transparency in media representation, giving viewers a direct signal when a person shown in an ad is not real. For marketers and creative teams, that shifts synthetic media from a production choice into a disclosure issue that has to be handled before an ad goes live.
The practical burden falls on advertisers and the agencies, studios, and vendors building synthetic creative for them. Teams using generated likenesses now need a repeatable way to identify where synthetic people appear, apply the required label, and document that review during approvals. The broader signal is that regulators are starting with disclosure rules first, rather than outright bans, but they are still moving synthetic media into a formal compliance framework.
- This creates a concrete compliance baseline for paid media that uses AI-generated people, which means disclosure checks need to sit alongside legal, brand, and platform review.
- Creative operations teams will need clearer asset-tracking and approval workflows so synthetic elements are identified early rather than discovered after campaigns are published.
- The law signals that synthetic media is no longer being treated as an edge case, and similar state-level labeling rules could spread to adjacent use cases beyond advertising.
Anthropic says it has taken its latest AI models offline to comply with new export controls
Anthropic said it took its latest AI models offline in response to new export controls, showing how quickly policy can interrupt access to frontier systems. The move is notable not just because it affects a leading model provider, but because it demonstrates that availability decisions can be driven by trade and national-security rules rather than technical readiness or product demand. For customers building on top of those models, access risk is now part of normal platform risk.
The decision also underlines a broader operational reality for AI vendors and enterprise users: model access can change with little warning when regulatory conditions shift. Teams relying on advanced external models may need to reassess dependency maps, fallback options, and regional deployment assumptions. In practice, export controls are no longer an abstract policy topic; they can directly affect uptime, roadmap timing, and customer commitments.
- Export controls can alter model availability suddenly for downstream users, so teams should treat regulatory disruption as a real service continuity scenario rather than a remote edge case.
- Vendor due diligence now needs to include geopolitical and trade-policy exposure, especially when a product depends on access to the newest frontier models.
- Engineering and product teams may need contingency plans for model substitution, staged rollback, or degraded-service operation if a primary provider becomes unavailable.