‘High’ priority to establish agentic payment protocols, says UK Treasury

Agentic AI is a term describing systems set up to act autonomously based on dynamic reasoning, with little or no human input. Use of agentic AI tools has been growing in popularity in the world of online shopping and payments, but the way agentic AI systems operate grates with existing payments regulations, as David Tilbury of Pinsent Masons highlighted earlier this month when exploring what new rules for agentic AI in payments might look like.

That fact has now been acknowledged by the Treasury: it said existing payment services regulations “may not fully facilitate the use of agentic AI”. It wants the UK to UK lead the world in agentic payments and has opened a consultation on modernising the regulations (43-page / 453KB PDF), giving industry the chance to have its say on how the framework should be adapted.

David Heffron of Pinsent Masons, an expert in financial services regulation, said: “The Treasury's consultation marks a decisive shift in the regulation of payment services in the UK.”

“Rather than simply updating existing rules, the government is proposing a more agile model in which detailed requirements increasingly sit in the FCA Handbook, while the statutory framework focuses on key principles, perimeter issues and consumer protections. Alongside this structural reform, the consultation seeks to future-proof regulation for stablecoin payments, tokenised deposits, agentic AI-powered payments and the next phase of open banking, signalling a clear ambition for the UK to become a global leader in payments innovation,” he said.

On agentic AI specifically, the Treasury said there is a “major opportunity for the UK to lead the development of agentic payments globally”. It said the UK regulatory framework for payment services and electronic money “must be set up to facilitate and support these innovations whilst managing risks to consumers and businesses”.

In its separate financial services AI adoption plan published on Tuesday, the Treasury said establishing a ‘trust framework’ to support agentic payments protocol is a “high” priority. It said agentic payments are a “near-term, practical proxy for a broader class of emerging autonomous financial systems”. As a result, it said getting the framework right for enabling agentic payments could pave the way for “broader, more complex agentic applications” across UK financial services.

Work in developing this new framework, the Treasury said, should be focused in three areas: on defining clear legal constructs and dispute mechanisms to unambiguously assign accountability when autonomous agents transact; establishing standardised identity and verification frameworks specifically designed for AI and autonomous software agents; and creating interoperable technical standards that ensure safe, frictionless, and trusted machine-to-machine authentication. It said industry should lead on the development of these new standards, but that government is open to legislating where appropriate “to provide clarity and enable safe adoption”.

Malcolm Dowden of Pinsent Masons, an expert in AI law, said: “The core issue in relation to payments made by agentic AI is determining when a payment effected by agentic AI exceeds authority given by the payer, whether through hallucination or through the goal-oriented behaviour of agentic AI. Focus will therefore be on the interaction between three provisions under the UK’s Payment Services Regulations 2017 – rules which establish consent or authority; rules requiring the payer to act within the issuer's terms and conditions; and rules that allocate liability for unauthorised payments.”

“A key problem with agentic AI is that it would first be necessary to work out whether the system has exceeded the authority conferred on it by the payer before being able to address allocation of liability,” he said.

In his Mansion House speech on Tuesday, Bank of England governor Andrew Bailey said that AI is likely to boost UK economic growth but warned that it also “creates problems of law”. He cited agentic AI as an example and said there is a need to address the questions its use poses “pro-actively and positively”.

Bailey said: “Typically, in law a principal becomes responsible for an agent while the agent operates within the remit given to it by the principal. If the agent goes beyond that remit, or acts improperly, it becomes responsible. Traditionally that had depended on the agent having a legal persona, i.e. being a human or a company etc. But when the agent has no legal persona does that leave the principal responsible at all times? If so, principals may need to constrain agents in ways that cut against the direction of innovation in AI models.”

“Put another way, is agentic AI in this context more a tool of the principal rather than strictly an agent, with the principal fully liable for all that is done and not done? Does this apply when real money is spent in ways that are final and binding, as and when agentic AI is allowed to hold the purse strings? This removes questions around the boundary of authority but should leave principals wondering how their governance frameworks can effectively operate with this technology. We must think about these things now: and I use this case study to illustrate why we must take a positive and pro-active approach to solving problems which are critical to ensuring effective innovation,” he said.

In her own speech at the Mansion House dinner, UK chancellor Rachel Reeves described AI as “the defining technology of our generation” and cited its importance to the UK economy and national security, highlighting the leading role the UK’s financial services sector has played in embracing AI – including in payments.

Last week, the Financial Conduct Authority (FCA) published the findings of a review into AI and the future of retail financial services, which was undertaken by FCA executive director Sheldon Mills. Mills predicted that agentic AI will change how people access and use financial services and how the financial services industry operates. Among his recommendations, Mills urged the FCA to carry out a review of the ‘regulatory perimeter’, to identify potential regulatory gaps that could arise from the technological shift he anticipates, and to adopt an “agentic supervisory model” in what would represent a material change in the way it addresses AI-related risk and regulates the industry.