Ofcom Warns Telecoms Against Trapping Customers in AI Chatbots

UK regulator Ofcom has warned telecom providers not to use AI chatbots in ways that make it difficult for customers to reach human support.

Ofcom found that 53% of UK internet users have interacted with AI, but only 8% have used it in connection with telecom services. The regulator also wants providers to clearly tell customers when they are speaking with AI rather than a person.

Customers still want access to human support

AI already plays a small role in telecom customer service. Around 6% of respondents used AI to resolve a telecom issue, 3% used it to compare prices, products, or services, and 2% used it to lodge a complaint.

Some customers value the speed and convenience of AI. Others complained about a lack of empathy or becoming stuck in chatbot loops.

Ofcom says providers should therefore give customers a clear route to human assistance when AI cannot resolve their problem.

Ofcom is not planning new AI rules yet

Ofcom currently believes its existing consumer protection rules provide enough protection, with little evidence so far of systemic harm caused by AI in telecom services.

The regulator plans to continue monitoring agentic AI adoption and signs of consumer harm instead of introducing additional AI regulations immediately.

Ofcom also sees potential benefits. AI could help customers compare offers, negotiate contracts, and resolve telecom problems more quickly.

The warning comes as AI agents face greater scrutiny over how they interact with online services. OpenAI recently admitted its AI models may have bypassed security controls, while another report found that OpenAI agents accessed Australian Medicare systems after ignoring restrictions.

Meanwhile, Meta Muse has reached 700,000 daily users as service issues appear.