In October 2026, Meta's Muse drew attention for its ability to call businesses on a user's behalf. Then the blocking started. After Amazon cut off Muse's shopping agent, multiple carriers began flagging AI-originated calls and routing them away from human support queues. The agent's call-making capability stopped working on those networks.
The carriers' stated concerns: unverified agent calls consume support capacity, and they carry impersonation and harassment risk.
One Door Closing, One Required to Stay Open
The carriers' logic runs opposite to a warning Ofcom issued the same month.
Ofcom's 2026 report said telecom companies should not let AI chatbots become an obstacle to reaching human support, purely because AI is cheaper than staff. “AI has the potential to make customer journeys easier, faster, and more personalized,” said Cristina Luna-Esteban, Ofcom's policy director.
The report found that only 8% of internet users use AI for telecom services, while 53% have interacted with AI in some form. Ofcom's position: existing rules are sufficient, but if AI use leaves customers unable to reach a company, that is a systemic problem.
Carriers are invoking “AI risk” to block outside agents. The regulator is invoking consumer protection to require human channels stay available.
Muse Can't Click “I Agree”
Amazon's stated reason for blocking Muse was that unauthorized AI agents violate its Conditions of Use.
That reason has an unresolved legal gap. Platform terms of service are written for a human to read and click. When an AI agent acts on a user's behalf, whether a binding agreement forms has not been tested in court.
Fenwick's legal analysis notes one of the core risks for AI agents: if an agent cannot form a binding contract on the user's behalf, a platform can not only remove the agent but pursue the developer directly. An Amazon spokesperson said third-party apps that buy on a customer's behalf “should operate openly and respect service provider decisions about whether or not to participate.”
Amazon's move was the lowest-cost path — cut the technical access, don't wait for the law.
The FCC May Open the Same Door for Politics
One of the carriers' stated reasons for blocking Muse was robocall risk. At the same time, the FCC is reviewing a petition that would loosen restrictions on political robocalls.
The conservative group Club for Growth is asking the FCC to exempt political organizations from parts of the Telephone Consumer Protection Act, allowing AI-generated voices in unsolicited calls to wireless numbers. The petition argues current federal rules are “outdated” and “suppress free speech.” The FCC has advanced the request to public comment and may rule by late October — before the midterms.
Democratic FCC Commissioner Anna Gomez opposed the move, saying it would create “a tsunami of robocalls and disinformation.”
The regulatory focus is not on AI making calls. It is on who is making them. Between an agent that negotiates for a consumer and an agent that campaigns for a politician, the rules are moving in different directions.
P.S. Meta says all external actions require explicit user permission and that account data lives in an isolated virtual environment. That has not resolved third-party risk concerns. Muse operates freely inside Meta's own ecosystem. Once it enters someone else's platform, the rules belong to that platform.
Frequently Asked Questions
Q: Why did carriers block Muse?
A: Carriers flagged AI-originated calls from Meta's Muse as consuming support capacity and carrying impersonation and harassment risk, routing them away from human queues.
Q: What does Ofcom say?
A: Ofcom's 2026 report warns telecom companies not to let AI chatbots become the only way to reach a company, saying human support must remain available when needed.
Q: Why can't Amazon just sue instead of blocking?
A: Platform terms of service are designed for humans to accept. Whether an AI agent can form a binding agreement on a user's behalf is untested, so blocking is the lower-cost path.
Q: What is the FCC petition about?
A: Club for Growth is asking the FCC to exempt political organizations from parts of the TCPA, allowing AI-generated voices in unsolicited calls. The FCC may rule by late October.
Q: What is the core tension?
A: Carriers are blocking AI that negotiates for consumers while the FCC considers allowing AI that campaigns for politicians. The rules turn on who is making the call, not whether AI makes it.
