UK bank blocked from cyber AI tool Mythos receives offer from rival OpenAI

UK bank blocked from cyber AI tool Mythos receives offer from rival OpenAI

OpenAI has provided nine major banks in the UK with access to its cybersecurity AI tool GPT-5.5 Cyber. That's because fierce competitor Anthropic blocked it from preview in its version, Claude Mythos.

Both tools are designed to find hidden security weaknesses in digital systems and can outperform humans in some hacking and cybersecurity tasks.

Mythos caused quite a stir when it was announced in April, with Anthropic claiming to have discovered a weakness in a legacy system that had been hidden for nearly 30 years.

The AI ​​Security Lab, which tested both Anthropic's Mythos and OpenAI's GPT-5.5 Cyber, said in its report that they reached "comparable levels of performance" on the tasks they set out.

Last week, Bank of England Governor Andrew Bailey warned that UK banks still do not have access to Mythos to test the security of their own digital systems and apps.

Former British Prime Minister George Osborne, now a senior executive at OpenAI, told the BBC that Bailey had never contacted him directly. He said his company "didn't want to hide [5.5 Cyber] or keep it to ourselves, but we made sure it wouldn't be available to everyone like Mythos."

“The key with these tools is that they have to be in the hands of the right people,” he said.

“We want to make sure that the forces that are trying to bring order to our democracy have these tools, and that the forces that are trying to disrupt us or commit crimes don’t have them.”

UK banks that now have access to OpenAI's GPT 5.5 Cyber ​​include Lloyds Banking Group, HSBC and Nationwide. NatWest and Santander already have access under existing contracts.

According to company officials, Antropic is urgently working to expand accessibility to Mythos. They believe that Mythos' features are in layers above 5.5 Cyber ​​and therefore require more attention.

Anthropic initially pitched it to a group of 42 companies, mostly other U.S. technology companies.

OpenAI has made access to Cyber ​​​​5.5 more generous, opening it up to banks in the EU, Japan, Canada and elsewhere.

Anthropic has invested $100 million in the Mythos preview, but both companies charge to use it.

Professor Alan Woodward, a cybersecurity expert at the University of Surrey, said: "We don't expect these tools to eventually surface something that humans wouldn't find, but they do catalog the millions of lines of code in banking apps alone with a ruthless thoroughness."

“In the UK, some of the code that the banking system is still running on is incredibly old, so AI will be very useful in finding problems there.

“Weeks of work can be done in minutes. “But we also get false positives, so we still need people to check the results.”