While Ofcom is alive to the risks that AI could pose to consumers, limited evidence of systemic harm has dissuaded it of the need for further regulatory intervention at this time
The proportion of consumers using AI for help with telecoms services is expected to increase rapidly
On 25 September 2026, Ofcom published a discussion paper outlining AI’s impact on telecoms customers so far and how this may evolve, alongside the short- and long-term regulatory issues that the technology’s adoption may raise. The paper found that customers in the UK are increasingly interacting with AI, with 48% of adult internet users having used generative AI tools, such as ChatGPT and Google Gemini, in the past 12 months. A smaller proportion (8%) used AI for help specifically with telecoms services, including to make a price or product comparison, or to lodge a complaint, although Ofcom suggests that this figure could increase quickly and will continue to track its growth going forwards. The report also found that there were mixed attitudes to AI, with many consumers feeling uncertain about whether to trust the outputs or results provided by AI chatbots. As AI continues to be deployed and used across the sector, Ofcom stated that it will maintain an outcomes-focused, technologically neutral approach, noting that some operators felt a more prescriptive approach to AI regulation could risk stifling innovation.
While AI is increasingly being used in customer service, there is a risk that the technology could act as a “gatekeeper” to human support
Across different cohorts of the UK population, Ofcom’s research pointed to concerns about the possibility of consumers being “forced” to interact with AI and not being able to speak to a human, especially in relation to complex or sensitive issues such as debt or bereavement. Ensuring that access to human support remains available alongside the deployment of AI tools was therefore highlighted as important, particularly for vulnerable customers or for those who may be unable or unwilling to engage with AI. Ofcom noted, however, that there were various benefits of using AI in customer service, such as a more personalised experience or faster response times, and that many telecoms operators already have guardrails in place to restrict AI-augmented chatbots to answering relatively simple questions. However, the regulator clarified existing rules under its General Conditions of Entitlement and its expectations of operators with respect to AI, including:
It should be made clear to telecoms customers when they are interacting with a generative AI application;
Operators remain responsible for any AI tools they deploy, even if these were built by a third party; and
AI applications should not become “gatekeepers” that create barriers to accessing human support.
Ofcom is not currently to impose additional rules to protect consumers from AI-related risks
While current deployments of AI mostly involve chatbots answering simple queries, Ofcom expects AI innovation to pose new regulatory challenges in the near future, particularly with regards to hyper-personalisation, one-stop AI intermediaries and agentic AI. These functions could lead to risks surrounding transparency, accountability for AI’s decisions and fairness, with vulnerable customers being more at risk. However, despite recommendations from consumer bodies Citizens Advice and Which? that Ofcom introduces new measures to require greater transparency about where operators are using AI, the regulator found limited evidence of systemic harm and considered that its existing rules enable it to protect customers effectively. Ofcom also considered that its current rules give firms sufficient certainty to invest and innovate responsibly with customer-facing AI, although it will continue to engage with industry and consumer groups and will “act quickly” if evidence of harm emerges. The regulator will also pay particular attention to whether AI applications are accessible and inclusive by design, how vulnerable customers experience AI applications and whether customers face risks of harm from fraud or new cyber threats. This decision not to introduce new rules is consistent with the broader international picture, with regulators so far largely holding back from direct interventions regarding AI’s role in the telecoms market.
