AI Safety Crisis: FTC Subpoenas OpenAI and Anthropic as Experts Demand Laws Before It's Too Late
Federal regulators have subpoenaed two leading AI companies over safety incidents, as experts warn that artificial intelligence is advancing faster than the rules meant to control it.
The Federal Trade Commission (FTC) has issued subpoenas to AI companies OpenAI and Anthropic as part of an investigation into safety incidents involving their artificial intelligence models, according to reports [1]. The action comes amid growing public concern about the risks posed by advanced AI systems [1].
Anthropic co-founder Jack Clark has urged lawmakers to create "common sense standards" for artificial intelligence, arguing that clear rules are needed to keep AI safe without slowing innovation [2]. Clark said smart regulation can protect the public while still allowing companies to build new tools, and that the focus should be on real risks rather than rules that block progress [2].
Microsoft CEO Satya Nadella has called for AI to have an "emergency brake," writing that it is time "to step back and assess the trust architecture" of AI—referring to the systems and rules that keep AI safe and reliable [3].
The push for oversight extends beyond Washington. Stanford's 2026 AI Safety Summit examined the dangers of AI agents—software that can act on its own—with experts discussing new ways to test AI systems and the difficulty of making them safer and more trustworthy [4]. Anthropic is also working with nonprofits METR and SecureBio to assess risks and safety practices, and is in early-stage talks with other organizations including Hugging Face [5].
In a notable move toward transparency, Anthropic CEO Dario Amodei has given METR "employee-like" access to the company's systems, allowing the independent research group to monitor its AI models from the inside [6]. The arrangement lets outside experts check AI systems for risks before they reach the public, reflecting growing pressure on AI companies to open their work to outside review [6].
However, experts caution that regulation will be difficult, pointing to two problems: the need for global cooperation and the fast pace at which AI is advancing [7].