The use of artificial intelligence agents and compliance with consumer law in the UK

Artificial intelligence is continuing to evolve and remains a buzzword in many businesses around the UK.

The UK Government’s recent ‘AI Skills for Life and Work: General Public Survey Findings’ found that 73% of the public have used AI in their day-to-day life (as at the date of the survey).

Despite developments in technology and increased use, there is currently no general statutory regulation of AI in the UK and there is no centralised AI regulator in place.

Instead, the UK relies on existing regulators and legislation to govern the use of AI in the UK.

This article focuses on the application of existing consumer law to the use of agentic AI systems in the UK.

What is agentic AI?

Agentic AI is an AI system that can operate to achieve particular objectives, often with the limited supervision or input. These systems can manage problem-solving tasks over time without human oversight.

Examples of agentic AI include Apple’s virtual assistant, Siri, and Tesla’s autopilot and Full Self-Driving technology, which has recently been approved for use in the Netherlands, the first country in Europe to permit this.

The benefits of using agentic AI systems can include enhanced productivity and efficiency, improved customer experiences and error reduction and consistency.

However, there are inevitably risks associated with such systems, and in particular, where they make decisions that it was not anticipated or intended that they make, with potentially harmful consequences.

How does existing law apply to agentic AI?

The Competition and Markets Authority recently published its paper on ‘Complying with consumer law when using AI agents’, aimed at providing guidance to businesses that have adopted the use of AI agents.

The Paper highlights the following points to assist with compliance:

• The same rules apply regardless of whether a human provides the services or an AI agent is used. Under consumer law, customers must be treated fairly. This principle applies regardless of whether a human or AI agent provides the services, and regardless of whether a third party has created the AI agent used.
• Inform customers if you use an AI agent in the provision of services. Transparency with customers on the use of AI agents is recommended. Under consumer law, consumers should be provided with relevant information to allow them to make informed decisions and should not be misled. If there is a chance that use of an AI agent may affect a customer’s decision to work with you, you should inform them.
• In general, AI agents learn from input and feedback they receive to improve their decisions or actions. Therefore, testing AI agents to ensure that they can respect customer’s rights, comply with legislative requirements and avoid misleading users is an important part of adopting agentic AI systems.

Through testing, you will be able to confirm that the systems work as intended, and ‘train’ the system appropriately before this is implemented in the real world with customers.

• Monitor performance. Check in to ensure that the AI agent you are using is working as intended, complying with the relevant consumer law requirements.

• Adapt and refine. If you encounter any issues with an AI agent or realise that it is not operating as intended, ensure that you act promptly to address any problems.

Whilst the use of AI is welcome and has the potential to create efficiencies, the onus sits with service providers to ensure that proper standards are upheld.

If you have any questions, please contact a member of our Corporate Team.