OpenAI Steps In with Cyber AI Offer as UK Banks Are Denied Access to Anthropic’s Mythos

OpenAI has extended an invitation to nine prominent UK banks to utilize its artificial intelligence-driven cybersecurity platform, GPT-5.5 Cyber. This strategic move comes as its primary competitor, Anthropic, has prevented these financial institutions from accessing previews of its own advanced tool, Claude Mythos. Both AI systems are engineered to identify latent vulnerabilities within digital infrastructure, demonstrating capabilities that can surpass human performance in specific hacking and security assessment scenarios.

The introduction of Mythos sparked significant debate when it was revealed in April. Anthropic had claimed the model successfully uncovered a security flaw in a legacy system that had gone unnoticed for nearly three decades. Consequently, several central bankers, finance ministers, and industry financiers have voiced grave apprehensions, fearing that such powerful models could potentially compromise the integrity of global financial networks.

According to a recent assessment by the AI Security Institute, which evaluated both Mythos and GPT-5.5 Cyber, the two technologies exhibited "a similar level of performance" against the benchmarks set for the test. Despite this parity, access remains restricted. Last week, Andrew Bailey, Governor of the Bank of England, highlighted that UK banks were still barred from using Mythos to audit their own digital applications and systems.

George Osborne, the former UK Chancellor and now a senior executive at OpenAI, addressed the situation in an interview with the BBC. He noted that Bailey had not reached out to him personally. Osborne emphasized that OpenAI had no intention of hoarding GPT-5.5 Cyber, stating, "We don’t want to hide [5.5 Cyber] away or keep it to ourselves." However, he clarified that availability would not be universal, asserting, "The key things with these tools is that they need to be in the hands of the right people. We want to make sure that the forces that are establishing order in our democracies have these tools, and the forces that want to disrupt us or commit crime, do not."

Under the new arrangement, Lloyds Banking Group, HSBC, and Nationwide will secure access to OpenAI’s tool. NatWest and Santander already possess access through prior agreements. Meanwhile, a representative for Anthropic indicated that the company is actively working to broaden access to Mythos. The firm maintains that Mythos operates on a higher capability tier than GPT-5.5 Cyber, necessitating a more cautious rollout strategy. Initially, Anthropic granted preview access to a consortium of 42 firms, predominantly comprising other US technology companies.

In contrast, OpenAI has adopted a broader approach, granting access to GPT-5.5 Cyber in the European Union, as well as to banking institutions in Japan and Canada, among other regions. While both companies charge for usage, Anthropic has invested $100 million to support previews of its Mythos tool.

Professor Alan Woodward, a cybersecurity specialist at Surrey University, provided context on the utility of these tools. "I don’t necessarily expect these tools to surface things humans wouldn’t find eventually, but they are relentless and incredibly thorough in sorting through the millions of lines of code which are in banking apps alone," he explained. He added that many UK banking systems rely on extremely outdated code, making AI particularly valuable for identifying issues. "They can do many weeks' worth of work in minutes – but they do also surface false positives, so you still need a human in the loop to check their findings."

Source: BBC News Generated at: 2026-06-01 17:07:20 UTC