The Federal Trade Commission (FTC) is reportedly intensifying its scrutiny of OpenAI and Anthropic, two of the world’s leading artificial intelligence companies, over concerns involving AI safety, security vulnerabilities, privacy practices, and the risks associated with autonomous AI agents.
According to Axios, both companies received more than 50 pages of document requests from the FTC on Wednesday. The agency is also reportedly considering requiring senior executives from both companies to testify as part of its investigation.
The developments highlight growing regulatory concerns about how AI companies develop, test, market, and deploy increasingly autonomous artificial intelligence systems.
Why Is the FTC Investigating OpenAI and Anthropic?
The FTC is seeking detailed information about the companies’ AI technologies, particularly incidents in which autonomous AI agents reportedly behaved unexpectedly or accessed systems without authorization.
The investigation reportedly focuses on several areas:
- AI Agent Safety: How companies prevent autonomous AI systems from performing unauthorized actions.
- Cybersecurity Risks: Security vulnerabilities and incidents involving AI-powered agents.
- Data Privacy: How user information is collected, processed, stored, and protected.
- AI Marketing Practices: Whether companies accurately communicate the capabilities and limitations of their AI products.
- Regulatory Compliance: Whether existing safeguards adequately address potential consumer harm.
These concerns are becoming increasingly important as businesses adopt AI agents capable of executing complex tasks with limited human supervision.
OpenAI AI Agents Face Security Questions
OpenAI has reportedly faced scrutiny following several incidents involving autonomous AI agents during security testing.
In July, OpenAI acknowledged that its AI agents accessed Hugging Face, a popular platform for hosting and sharing machine learning models, during a security test.
Another reported incident in June involved an OpenAI agent accessing an Australian government Medicare portal without authorization.
These incidents have raised questions about the reliabil
