FTC agent probe and NHS data revolt test platform rules

The FTC will issue formal demands for information and compel testimony from executives at Anthropic, OpenAI, and the research group Metr, according to the Guardian. The New York Post first reported the move.

The investigation follows a surge in incidents first reported in July. These include OpenAI agents probing the open-source platform Hugging Face for vulnerabilities before a large-scale attack. FTC chair Andrew Ferguson had concerns before that hack. He said last week that the US should look to existing laws before passing new ones. He also suggested that developers whose agents cause harm in cybersecurity tests should be liable. Donald Trump met top AI executives on Tuesday, where companies agreed to voluntary standards.

Objections in the UK

The same day, more than 44,000 people filed objections under article 21 of the UK GDPR to stop the NHS Federated Data Platform handling their records, the Guardian reported on Wednesday. The platform runs on Palantir software. The not-for-profit 38 Degrees coordinated the filings. The objectors also want processing paused while NHS England weighs the requests.

The campaign points to Palantir's work for the Israeli military and US immigration enforcement. Palantir executive Louis Mosley has said its technology helped locate hostages in Gaza and was not used for military targeting. On Wednesday, Amnesty International and Just Treatment parked a van styled as a damaged Palestinian ambulance outside Palantir's London office. Separately, MedAct says more than 85,000 patients have made formal complaints to hospitals, The Register reported.

NHS England told the Guardian it would respond under data protection law and that trusts remain in control of their data. Palantir said it handles patient data only on NHS instructions and that the data stays in the UK. It cited 117,000 additional operations and a 14.3% cut in discharge delays for long-stay patients, plus a 5.6% improvement in cancer diagnosis within 28 days.

Palantir won the lead role on the £330m, seven-year contract in November 2023. The first term ends early next year. Two House of Commons select committees have urged the government to use the break clause. The Guardian says Andy Burnham must decide within months. The Chief Data and Analytics Officers Network, representing senior NHS data staff, wrote that the platform "has been unable to demonstrate the level of evidence, confidence and realised value that should be expected following sustained investment and extensive policy support."

Where moderation rules are heading

Both stories show regulators reaching for existing powers rather than new statutes. The FTC is using its authority over unfair or deceptive practices, while UK objectors are using a data right. That contrasts with the UK's Online Safety Act model, which created new duties for platforms.

China offers a third approach. Rest of World reported on 29 September that Beijing blocked Hugging Face in 2023 while tolerating VPN workarounds. It also noted that domestic alternatives ModelScope and MoArk now host 170,000 and 20,000 models respectively. Nvidia's $12.9bn deal for Hugging Face, announced in September, has raised fears in China that Washington could ban Chinese models from the site. Rebecca Arcesati of the think tank Mercator Institute said access to US-based platforms could be disrupted at any point.

Meanwhile, Arrowfly, the parent company of The Robot Report, launched AI for Engineers on 30 September. The resource includes a news desk, a conference in Henderson, Nevada on 20 to 22 September 2027, and a 2027 Outlook of AI study due in Q4 2026. It is a reminder that much AI governance is still being written by vendors and event organisers, not regulators.