UK banking is at a divide. Anthropic’s Claude Mythos, an AI model designed to identify and patch software vulnerabilities, has been rolled out to roughly 200 partner organizations as part of ‘Project Glasswing’, but UK banks have so far been notably absent from the list.
Bank of England Governor Andrew Bailey has expressed serious security concerns about the gap, while several leading UK lenders are turning to OpenAI’s rival cyber-focused model instead.
For now, the UK sector is split between those racing to secure frontier AI defenses and those still waiting to do so.
Latest Videos FromTechRadar
Financial Services Specialist at Smart Communications.
It’s a genuinely important story for cybersecurity teams, but for everyday customers, it’s largely invisible. Nobody checks which AI model their bank has access to before deciding whether to trust it.
What they notice is whether their bank tells them what’s going on when something goes wrong. That’s a test that won’t go away, regardless of which side of the access gap a financial institution sits on.
In fact, consumer confidence in financial institutions is declining, with over a third of Brits distrustful of their providers. In this context, even the smallest incidents can escalate into reputational problems if communication is delayed or unclear.
Faster detection is changing customer expectations
Every major technology shift introduces new risks, but AI changes the speed at which those risks emerge and spread. Tools across fraud monitoring, compliance and cybersecurity are clearly getting faster at surfacing problems. That’s good news operationally, but it widens the gap between how quickly a bank knows about an issue and how quickly it tells its customers.
If a bank’s systems can identify suspicious account behavior in real time, customers reasonably expect their bank to communicate in real time too. Response time is a big trust factor among UK consumers. If systems like Mythos can recognize emerging issues immediately, customers expect guidance around the same issues immediately too.
With people unsure of whether to trust AI, silence from a bank creates its own narrative. Customers begin to fill information gaps themselves, often through online speculation or fragmented reporting elsewhere. In financial services, trust is rarely lost in a single catastrophic moment. More often, it weakens through confusion and perceived inaction.
Communication is now part of risk management
This is why communication can no longer sit downstream from operational infrastructure. It must become part of risk management itself.
When incidents occur, customers want clarity about what is happening, confidence that the institution understands the issue, and clear guidance on what they need to do next. Too often, businesses focus on technical remediation, while customer communication is delayed by fragmented systems or uncertainty about messaging.
That fragmentation is already evident in day-to-day experiences, with only 59% of financial services consumers saying their provider communicates using their preferred channel. As a result, most people receive updates too slowly, across disconnected channels, or in language that prioritizes compliance over clarity.
That approach no longer works in environments shaped by AI-speed detection and real-time digital expectations. The institutions that maintain trust most effectively will be those that communicate with the same precision and responsiveness as the systems that monitor the risk itself.
So, what are the top ways banks can maintain trust?
Prioritize speed alongside accuracy – in high-pressure moments, delayed communication often creates more customer anxiety than the issue itself.
That does not mean rushing incomplete information to customers. But it does mean building communication workflows that enable organizations to quickly acknowledge issues, explain what is known, and provide regular updates – on the customer’s preferred channel – as situations develop. Customers are generally more understanding of operational problems when they feel informed throughout the process.
Use plain language instead of technical reassurance – many financial incidents become harder for customers to process because explanations are overly technical or overly cautious.
Customers do not want pages of operational detail. They want straightforward explanations of what happened, whether they are affected, and what action they should take. Clear communication reduces uncertainty. It also helps prevent panic-driven behavior that can escalate customer support pressure during incidents.
As AI systems become more complex behind the scenes, communication needs to become simpler at the customer level, not more complicated.
Deliver consistency across every channel – customers move between apps, email, SMS, websites and contact centers interchangeably, and they notice when the message varies depending on where they look.
One of the biggest risks during fast-moving incidents is fragmentation: customers seeing conflicting guidance depending on where they look or the channel that they are on. Maintaining trust requires coordinated communication that delivers the same message, tone and guidance consistently across every channel and customer interaction.
Trust will be won or lost through communication, not the AI model
Mythos will not be the last AI system to reshape financial services operations. Technologies across fraud prevention, compliance monitoring, customer support and risk analysis are already impacting how institutions operate and communicate with people.
But operational capability alone will not determine which organisations succeed. As systems move faster and risks surface earlier, trust increasingly depends on how institutions communicate during moments of uncertainty.
Customers don’t expect perfection. They understand that new technologies introduce complexity and evolving risks. What they expect is transparency, responsiveness and reassurance when issues arise.
It’s about whether the behind-the-scenes infrastructure can turn detection into the right message, on the right channel, fast enough for it to land before doubt takes hold.
We’ve listed the best patch management software tools.
This article was produced as part of TechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.
The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/pro/perspectives-how-to-submit
https://cdn.mos.cms.futurecdn.net/xHBS5bqnPRLvifkuzjdknM-2560-80.jpg
Source link
