FCA Emerging Technology Horizon Scan 2026 – Tomorrow’s World?

The Emerging Technology Horizon Scan 2026 issued on 10 June 2026 is described as the FCA's first external publication of its kind. The paper is not regulatory guidance or predictions, but a forward-looking view based on research and analysis by the FCA's Emerging Tech & Research team on how new technologies such as AI and distributed ledger technology (DLT) could reshape financial services in the next few years. The paper goes from a ‘state of the world’ chapter to three plausible scenarios: personalised intelligence; synthetic (in)security; and programmable finance. These are considered ‘plausible’ and are grounded in reality. For each scenario, the FCA describes it, discusses the route to 2030, and posits some conclusions.

There are a few key takeaways from the report, including:

• the same technologies enabling efficiency, improved financial decision-making and inclusion simultaneously increase risk and potentially expand the toolkit of bad actors; and
• the landscape is set to shift significantly in response to emerging technologies; if the FCA has already identified plausible scenarios, it is reasonable that firms should also be thinking about forthcoming change and planning how they will respond.

Scenario 1: Personalised Intelligence

The report foresees consumers increasingly delegating financial decisions to AI agents that anticipate their needs and act on their behalf — sometimes without the consumer noticing. Ever-increasing data collection will drive not just product recommendations but the personalisation of products themselves. On the bright side, this could improve access, including for consumers with ‘lower financial dexterity or awareness’, reduce inertia in switching, and help consumers avoid bad outcomes and scams. AI proxies could act as cognitive extensions, continuously learning from behavioural and biometric signals to optimise financial decisions in real time.

However, the report flags significant risks: over-delegation, cognitive erosion and psychological vulnerability. Consumers may accept their AI proxies' prompts for consent with the same casualness they apply to web cookies today. Hyper-personalised products may become too complex for consumers to understand or firms to explain, and a two-tier market of ‘human + AI’ versus ‘human-alone’ consumers could deepen inequalities. And if the AI proxies are designed or funded by firms with commercial interests, they may introduce subtle biases favouring certain providers without the consumer realising.

The report also warns that new ‘dark patterns’ could emerge that target AI proxies rather than people — for example, mis-selling may no longer occur primarily through persuasive sales scripts directed at humans, but through adversarial optimisation intended to pass a proxy’s filters while still meeting the letter of formal disclosure requirements. 

Also on the negative side, the speed at which agentic systems operate means that they can execute thousands of transactions or decisions in seconds. If these are misaligned or manipulated, the velocity of harm outpaces traditional firm oversight, making detection and remediation difficult. 

Scenario 2: Synthetic (in)security

The FCA's report anticipates that ‘AI-fication’ of human thought, labour and digital infrastructure will create a future in which simulated data becomes indistinguishable from real data, and where a single human prompt could orchestrate a global network of synthetic scammers. According to the report, we are entering an era where evidence no longer guarantees authenticity, and where the primary indicator of criminal activity may be suspicious perfection rather than anomaly. But the report stresses that the AI-powered financial crime threat does not stop there:

Synthetic crime and the democratisation of fraud: The report highlights that the commoditisation of AI is lowering barriers to entry for sophisticated crime; where large-scale fraud previously demanded networks of people and specialist expertise, a single individual can now deploy thousands of AI agents at minimal cost. The rise of Crime-as-a-Service means less-skilled criminals can rent AI tools on the dark web to execute attack campaigns rivalling those of large cybercrime groups. Synthetic systems can fabricate not just data but entire ecosystems of plausibility, eroding the evidential integrity on which markets depend.

Deepfakes and cognitive warfare: The report notes that synthetic deception is moving beyond realistic images and audio into ‘cognitive warfare’ aimed at influencing not just what people see, but how they think and decide what to trust. The cognitive bandwidth of consumers, regulators and financial professionals becomes a critical attacksurface and traditional approaches to verification may no longer be effective.

Narrative laundering and synthetic evidence: According to the report, AI may generate complex synthetic evidence trails (e.g. fabricated audit trails, negotiation emails, shipping manifests and transaction histories) to conceal financial crime such as Ponzi schemes, bribery or terrorist financing. This could make misconduct structurally invisible to both firms’ controls and supervisory tools.

Synthetic market abuse: The report warns that contemporary systems may fail to recognise new forms of market manipulation emerging from the unintended behaviours of multi-agent systems, which may commit insider trading, collusion, spoofing or pump-and-dump strategies without human involvement in the execution. In high-stakes adversarial trading markets, agents could manipulate markets by executing strategies that human analysts cannot detect. A ‘swarm’ of AI agents could coordinate to manipulate markets by manufacturing a false reality around a financial entity, creating a synthetic ‘consensus cascade’ which poses a direct threat to the informational integrity on which market confidence depends.

Expanding attack surfaces and adaptive threats to firms’ operational resilience: The report identifies that frontier AI models can now discover ‘zero-day’ vulnerabilities (i.e. newly discovered vulnerabilities without a fix or mitigation available) across the financial services ecosystem in record time. In the wrong hands, such technology can cause serious harm and reduce the ability of security teams to respond effectively. Another major threat to operational resilience highlighted by the report is adaptive malware - AI-powered software that can adapt and change in real time once it has infected the host. Such malware can alter the way it attacks, rewrite its own code to evade detection and imitate legitimate activity to hide what it is doing. A novel spin on operational resilience is the idea of 'cognitive reliance', where people become so dependent on AI tools and automation that they are no longer able to act for themselves without those systems and so become vulnerable in the event of failure. This risk is exacerbated as the largest AI providers become ubiquitous across financial services.

Scenario 3: Programmable finance

The FCA's report looks at UK growth opportunities arising from the convergence of DLT with other financial concepts to deliver new types of financial instruments, infrastructure and systems. Decentralised protocols, digital assets and smart contracts have significant implications for the future structure of financial services, existing alongside the centralised platform economy. As the FCA paper observes: ‘The sharp distinction that once existed between traditional finance (TradFi) and [decentralised finance] is beginning to fade.’

In terms of what this could look like in the near term, the FCA paper identifies the following:

Sovereign digital ecosystems: Responding to different priorities, jurisdictions will build modular digital financial ecosystems which ‘leverage converging technologies to bring identity, data and payments into integrated national foundations’. As diversity in approaches expands, the FCA's report foresees that ‘strategic opportunity’ for jurisdictions may sit with how ‘effectively they build the technical, legal and regulatory bridges that enable value to flow seamlessly across divergent financial stacks’. 

A shift from faster automation to autonomy: This envisages a progression to seamlessness. The paper offers an example of a business hardwiring compliance into its logic layer with an instruction to pay a supplier only when real-time data confirms that the product in a shipping container has cleared a port. In this example, the trade, insurance and payment are ‘collapsed’ into a single atomic event rather than being distinct stages.

Built-in compliance and the ‘finternet’ vision: Here the FCA considers how in the future ‘intelligence, compliance and control’ may be embedded in financial services via the mechanisms of shared ledgers and synchronisation tools that can ensure settlement across tokenised assets and money or stablecoins without reconciliation risk.

Conclusion

This is a new type of communication from the FCA and one to be welcomed. The point is not whether this vision of tomorrow’s world will necessarily come to pass, rather the presentation of plausible scenarios allows consideration and engagement from regulated firms, fintechs, providers to the industry and the public. For regulated firms in particular, the report presents some challenging ideas about the fundamental changes which the future may bring and the need to grapple both with the positive benefits and the risks of harm.

Aside from this paper, it will be interesting to see more generally how regulators (and firms) tackle these risks and in particular, whether we will see a change of tack from the UK regulators’ long held ‘technology neutral’ approach to rule making. It certainly seems more in keeping with the technology-positive stance from the Government; its push for growth leveraging technology is evident from the AI Opportunities Action plan dated January 2025, which set out objectives for the increased use of AI to drive economic growth and open up new opportunities, and the Chancellor’s speech at the AI Adoption Summit in June 2026 in which she announced further initiatives in relation to this technology. 

Another facet of the tech-neutral approach is the absence of specific rules relating to technology or the mode of delivery, with the FCA increasingly focused on outcomes rather than how those outcomes are delivered. In contrast, other jurisdictions have been more prescriptive in regulating AI (e.g. under the EU AI Act, albeit currently being reconsidered under the AI Continent Action Plan) and firms operating across borders will face a complex, multi-standard environment in this area.