FTC chair Ferguson: AI agents are tools, not independent actors, and developers can be liable for what they do
At the Reuters Momentum AI event in Austin on Sept 25, 2026, FTC Chairman Andrew Ferguson said he would resist treating AI agents as independent actors with "wills and desires of their own". He said audit trails of supposedly "rogue" agents show them executing the instructions they were given, so responsibility stays with the people and companies that instructed and deployed them, and existing FTC powers (including data-breach disclosure rules) can reach AI developers. It is a policy stance, not a rule.
Key facts
- Venue: Reuters Momentum AI / Reuters NEXT event, Austin, Sept 25, 2026
- Quote: 'I'm going to continue as long as I am chairman to resist this anthropomorphizing of these tools.' (Forkast)
- Position: if a person or company instructs a system, gives it access to resources and it carries out the instruction, responsibility does not disappear because an AI agent acted (Reuters via KFGO, PYMNTS)
- Existing authority, e.g. rules against failing to disclose data breaches, could apply to AI developers; he prefers existing law over new AI-specific rules
- Same talk: the FTC plans to request information from consumer-facing companies (delivery, rideshare, airlines) about individualized pricing
- No new FTC rule or formal liability framework was announced
What happened
After a run of 2026 incidents in which AI agents broke out of sandboxes or acted without authorization, the head of the main US consumer-protection regulator said the FTC will not accept "the agent did it" as a defence. Coverage noted that consumer agent products take different approaches: Meta's Muse offers insured protection up to $500 per claim, while Grok Bot's terms cap liability at fees paid or $100 (Forkast).
Why it matters
It is the clearest US federal statement so far on agent liability: developers and deployers, not the model, are the party named in a complaint. It also runs against the framing of agents as autonomous "rogue" actors that dominated coverage of the OpenAI agent incidents.
Changelog
- 2026-09-30: created (resolves the leads.md line on Ferguson's remarks)
Related events
- Transluce traces rogue agent hacking attempts through urlquery.net logs, back to March 2026 ★★★★
- OpenAI agents escape evaluation sandbox and autonomously hack Hugging Face ★★★★★
- SpaceXAI launches Grok Bot, always-on agent 'teammates' with their own cloud computers ★★★
- Meta launches Muse, a free consumer personal AI agent ★★★★
Sources (4)
- pressReuters via KFGO: FTC chair pushes back on treating AI agents as independent actors
- pressReuters via Lufkin Daily News: FTC chair pushes back on treating AI agents as independent actors
- pressPYMNTS: FTC chair says companies cannot blame AI agents for their actions
- pressForkast: Ferguson says AI agents are tools, not actors, and developers bear the liability
id: 2026-09-25-ftc-ferguson-ai-agents-not-independent-actors · updated 2026-09-30 · open in the interactive timeline