OpenAI has joined the Global Anti-Scam Alliance as a foundation member, a step the two organisations announced from The Hague. The alliance is a coalition of companies, regulators and researchers working against fraud; the foundation tier is its highest form of membership.

OpenAI says it already invests in detecting and disrupting attempts to use its systems for scams, phishing and fraud, and that it publishes lessons from that work. Under the membership it will contribute to the alliance's research, sit on its advisory board, and take part in cross-sector efforts to reduce the harm caused by AI-enabled fraud.

Sharing lessons is not the same as sharing data

A foundation membership is a commitment to cooperate, and cooperation is worth having. But the announcement says nothing about what will actually move between the parties: whether scam telemetry and blocked-attempt patterns will be pooled, or only the published write-ups that already exist. That distinction decides whether the alliance changes anything or simply adds a logo to a page.

There is also a commercial tension that neither organisation addresses. The most useful intelligence about AI-enabled fraud sits inside the platforms where it happens, and publishing enough of it to help defenders also exposes how the tools are being abused. Will McCants, who leads intelligence and investigations at OpenAI, said the company's tools are getting better at spotting scams; Jorij Abraham, the alliance's managing director, said coordination across technology, policy and law enforcement is what the problem demands. Both are reasonable. The question a reader should keep is simpler: when the alliance reports progress, what number will it publish, who will be able to check it, and whether any of it comes from outside the members' own press releases?