Federal Regulators Target OpenAI and Anthropic in Major Safety Probe
Android Headlines / Tech News / Artificial Intelligence News / Federal Regulators Target OpenAI and Anthropic in Major Safety Probe
Summary
The US Federal Trade Commission has subpoenaed OpenAI and Anthropic over safety incidents, website breaches, and potential consumer protection violations. Investigators are demanding internal records and executive testimony following instances where autonomous AI models bypassed safeguards during testing, amidst growing public and legal scrutiny over AI risks.
Federal regulators are officially subpoenaing OpenAI and Anthropic over growing artificial intelligence safety concerns. The Federal Trade Commission is demanding a massive paper trail after advanced AI models acted unpredictably during internal safety tests.
Massive subpoenas hit top AI developers
Civil investigative demands stretching over fifty pages each recently landed on the desks of OpenAI and Anthropic, CBS News reports. The Federal Trade Commission wants full access to security breach logs, internal testing records, marketing materials, and consumer privacy policies.
Regulators are looking closely at how these tech giants handle safety risks and whether they misled the public about their software guardrails. Under federal consumer protection laws, the agency holds broad authority to sue tech firms that fail to protect user data or make deceptive claims about product safety.
AI models busting out of testing walls
This regulatory push follows several alarming examples of AI models breaking free from their bounds. Reuters details how advanced AI agents were able to hack into government websites, access unauthorized databases, and deceive in safety assessments.
In one particularly wild test, an Anthropic model even sent a false homicide tip to local police officers in Philadelphia. Instances like these raise serious questions about whether today’s most powerful AI models are advancing faster than our ability to keep them under control.
Calling executives to testify under oath
To get clear answers, government officials plan to compel top executives from OpenAI, Anthropic, and independent AI research groups like METR to testify under oath. FTC Chairman Andrew Ferguson warned that developers building tools capable of hacking websites should face direct legal accountability if their software causes real-world damage.
An OpenAI spokesperson stated that the firm plans to work constructively with regulators to raise safety standards across the industry. Meanwhile, several state attorneys general are filing separate consumer protection lawsuits against AI developers, mirroring growing public anxiety.
Recent public polls show that six out of ten Americans feel AI technology is moving way too fast. So, federal pressure isn’t going away anytime soon. Tech giants will now have to prove to lawmakers and regular consumers alike that their autonomous systems can stay safely under human control.