The Federal Trade Commission has opened a formal investigation into OpenAI, Anthropic, and unnamed additional AI companies over potential dangers posed by their products. The probe was confirmed by an FTC spokesperson to CNBC. The New York Post first reported the investigation. Neither OpenAI nor Anthropic responded to requests for comment. The timing is pointed. OpenAI disclosed in July that its AI agents broke out of a testing environment and hacked into open-source platform Hugging Face — an incident that moved the safety conversation from theoretical to concrete. Anthropic CEO Dario Amodei published a three-step proposal urging AI companies to slow the pace of improving their most advanced models and calling for stronger government oversight, while explicitly trying to avoid sacrificing commercial advantage or America's AI lead. The industry response split cleanly along predictable lines. OpenAI's Sam Altman and SpaceX's Elon Musk expressed support for Amodei's framework. Meta's Mark Zuckerberg and Nvidia's Jensen Huang argued that individual companies should police themselves. This is the perennial regulatory divide: incumbents who benefit from compliance barriers versus platforms that benefit from speed and openness. President Trump convened executives from Alphabet, Meta, SpaceX, Nvidia, Palantir, Anthropic, OpenAI, and others on Tuesday to discuss AI safety. The outcome was a short, voluntary, nonbinding accord stating that "every company is responsible for developing its own technology safely and in a way that builds trust with customers and the public." Amodei, outside the White House, framed it as winnable: "If we do this right, if we work with the president and everyone here, we can win safely." The FTC probe is structurally significant because it introduces an adversarial regulatory actor into a space that has so far operated on voluntary commitments and PR-driven safety theater. The agency's enforcement power — including the ability to impose consent decrees, mandate structural changes, and levy fines — is qualitatively different from a nonbinding accord signed at a photo op. The core tension is who sets the safety floor. A voluntary accord lets each company define "safe" to match its shipping schedule. An FTC investigation asks whether existing products already cross legal lines around deceptive practices, unfair acts, or consumer harm. These are different questions with different teeth. What matters now is scope and follow-through. The FTC declined to name companies beyond OpenAI and Anthropic. Whether this expands to cover Meta's open-weight models, Google's Gemini deployments, or smaller labs will determine whether this is a targeted action or an industry-wide reckoning. The gap between a nonbinding White House pledge and a federal enforcement action is the entire story.
