UK consumers lost more money to financial fraud in 2025 than at any point in the past four years, according to new figures that are already reshaping the debate between banks, regulators and the technology platforms that increasingly sit at the centre of scam ecosystems.
The headline number—losses rising again after a period of attention and investment—matters not just because it signals a worsening trend, but because it intensifies a specific political pressure point: banks want the government to require tech companies to do more to prevent fraudulent activity from spreading at scale. In practice, that means moving beyond voluntary cooperation and towards clearer accountability for platforms, messaging services and other digital intermediaries that scammers rely on.
For consumers, the impact is immediate and personal. Fraud is rarely a single event; it is usually a sequence—first contact, then persuasion, then exploitation, often followed by attempts to keep victims from reporting or reversing payments. The new data suggests that this sequence is becoming harder to interrupt, even as banks improve detection systems and strengthen customer protections.
What makes the 2025 figures particularly consequential is the way they frame responsibility. Traditional discussions about fraud have tended to focus on banks’ controls: authentication, monitoring, transaction limits, and the speed at which suspicious activity is flagged. But the latest push is different. It argues that the “front door” of many scams is now outside the banking perimeter—on social media, in messaging channels, through search results, or via advertising networks that can be exploited quickly and repeatedly.
That shift is forcing banks to ask a broader question: if fraudsters can recruit victims through platforms that are designed for rapid reach and engagement, why should the burden of prevention fall primarily on financial institutions?
A market that rewards speed—and punishes friction
Scammers operate with a kind of industrial logic. They test messages, refine scripts, rotate accounts, and move quickly when takedowns occur. The digital environment is built for speed: content spreads instantly, accounts can be created rapidly, and communications can be routed through multiple services before anyone notices. Even when platforms remove obvious scam content, fraudsters often adapt within hours—changing wording, using new accounts, shifting to different channels, or targeting different demographics.
Banks, by contrast, are constrained by regulation, risk frameworks and the need to avoid blocking legitimate customers. That creates a structural tension. If banks become too aggressive, they risk false positives that inconvenience genuine users. If they become too cautious, fraudsters exploit the window between suspicious behaviour and intervention.
The result is a cat-and-mouse dynamic. The 2025 rise in losses suggests that, despite ongoing improvements, the system still isn’t closing the loop fast enough—especially when scams are coordinated across multiple platforms and payment rails.
Why “platform accountability” is gaining traction
The argument for greater accountability for platform and technology firms is gaining momentum because it aligns with how modern scams are actually executed. Many fraud journeys begin with a lure: a fake investment opportunity, a counterfeit customer service interaction, a romance scam that escalates into requests for money, or a “delivery problem” message that leads victims to a fraudulent payment page. The early stages often involve content distribution and identity manipulation—areas where banks have limited visibility.
Banks can monitor transactions once money is moving. But they cannot easily see the earlier persuasion mechanics: the ads that convinced a victim, the group chat that built trust, the impersonation profile that established credibility, or the link that redirected a user to a scam site.
This is where policymakers and banks increasingly want tech companies to step in. Not necessarily by policing every piece of content in real time, but by improving the operational capabilities that reduce scam throughput. That includes faster detection of coordinated fraud campaigns, better enforcement against repeat offenders, stronger identity and account integrity measures, and more effective takedown processes that don’t simply remove one post while leaving the underlying infrastructure intact.
The unique challenge is that scams are not always “static.” They evolve. A platform might remove a specific advertisement, but the scam network can immediately reappear under a new account or a slightly altered message. That’s why banks and consumer advocates are pushing for accountability that focuses on systems and outcomes—not just individual removals.
From detection to disruption: what banks are likely to prioritise next
As fraud continues to evolve, the next phase of bank-led strategy is expected to emphasise disruption rather than only detection. Detection catches suspicious patterns after they emerge; disruption aims to stop the scam journey earlier, ideally before funds are irreversibly transferred.
Several areas are likely to receive renewed focus:
First, faster detection and response. This doesn’t only mean better algorithms; it also means operational readiness—how quickly alerts are triaged, how quickly teams can intervene, and how quickly customers can be contacted in a way that is both effective and compliant. In many cases, the difference between a prevented loss and a completed scam is measured in minutes.
Second, stronger take-down processes. Banks can’t take down content themselves, but they can push for improved coordination with platforms and for clearer escalation pathways. The goal is to reduce the time between identification of a scam campaign and its removal across channels.
Third, improved consumer support. Fraud prevention is not only technical; it is behavioural and emotional. Victims often feel shame, confusion or fear of blame, which can delay reporting. Better support—clearer guidance on what to do immediately, reassurance that reporting won’t automatically lead to punitive outcomes, and faster pathways to recovery—can reduce the overall harm even when prevention fails.
Fourth, tighter feedback loops between banks and platforms. If banks can share indicators of scam infrastructure—such as domains, phone numbers, account identifiers, or behavioural patterns—platforms can act more quickly. But sharing must be structured and actionable, not vague. The push for government involvement often reflects frustration that voluntary information exchange has been inconsistent.
The human cost behind the statistics
Fraud losses are often discussed in aggregate terms, but the lived experience is more complex. Victims may be targeted through multiple channels: a message that appears to come from a trusted organisation, a call that mimics a bank’s tone and branding, or a “verification” request that leads to credential theft. Once a victim is engaged, scammers frequently escalate pressure—claiming urgency, threatening account closure, or insisting that the victim must act immediately to avoid consequences.
This is why the 2025 increase is so alarming. It suggests that the scam playbooks are working well enough to overcome improvements in bank controls. It also implies that the ecosystem around fraud—platform reach, identity deception, and payment facilitation—remains fertile.
In many cases, victims are not simply tricked once. They are guided through steps that make them complicit in their own loss: transferring funds, authorising payments, or providing access to accounts. That makes recovery harder, because the transaction may appear legitimate from a purely technical perspective.
The policy implication is clear: if fraud is increasingly a multi-stage process involving platforms and communications services, then prevention must be multi-stage too.
What government action could look like
The pressure on government is not just rhetorical. Banks and consumer groups are effectively asking for a framework that forces tech companies to treat fraud mitigation as a core responsibility rather than an optional goodwill measure.
While the exact shape of such requirements can vary, the direction implied by the current debate typically includes:
Clearer obligations for platforms to address fraud-related harms, including measurable standards for response times and enforcement effectiveness.
More robust account integrity requirements, such as stronger verification for high-risk activities and better controls against automated creation of scam accounts.
Improved transparency around enforcement actions, so that regulators and partners can understand whether interventions are working.
Stronger mechanisms for rapid takedown and escalation when fraud campaigns are identified.
Better coordination across sectors, potentially supported by government-backed processes that reduce friction between banks, platforms and law enforcement.
The key is that accountability must be operational. A requirement that simply says “remove fraudulent content” is unlikely to solve the problem if scammers can reappear instantly. Requirements that focus on reducing scam throughput—by disrupting infrastructure, limiting repeat abuse, and improving detection of coordinated campaigns—are more likely to produce measurable reductions in losses.
A “scale problem” rather than a “single incident” problem
One reason the 2025 figures are hitting hard is that they highlight a scale problem. Fraud is not limited by geography; it is limited by attention and access. Platforms provide both. When scam content can be distributed widely and quickly, fraud becomes less like isolated crime and more like a scalable business model.
That business model depends on three things: visibility (getting in front of victims), credibility (making scams look legitimate), and payment capability (moving money out of reach). Banks can influence the third element, and sometimes the second through customer education and authentication. But the first element—visibility—is largely controlled by platforms and advertising ecosystems.
So when losses rise, banks interpret it as evidence that the upstream pipeline remains too open. That is why the debate is shifting from “what banks can do” to “what the broader digital environment must do.”
The role of AI and automation in both fraud and defence
Although the immediate story is about fraud losses and accountability, the broader context is the growing role of automation and AI in scam operations. Fraudsters use AI to generate convincing messages, personalise lures, and mimic customer service interactions. They can also automate account creation and content posting, increasing volume and reducing the time between attempts.
Defenders are also using AI, but the challenge is that fraud and defence are racing each other. A detection model can be trained on known patterns, but scammers constantly change tactics. That means defence systems must be adaptive and integrated into real-time workflows, not just deployed as static tools.
This is another reason platform accountability matters. If fraud campaigns are detected but not disrupted quickly across channels, the attackers can simply pivot. Effective defence requires coordinated action across the entire chain—from content distribution to payment processing.
What consumers can do right now (and what banks will likely encourage)
Even as policy debates continue, consumers remain the frontline. The most effective advice tends to be simple, but it must be repeated because scams