In a significant development for UK financial institutions, OpenAI has now granted access to its advanced cybersecurity AI tool, GPT-5.5 Cyber, to nine major banks. This comes on the heels of fierce competitor Anthropic deciding to block previews of its own cutting-edge tool, Claude Mythos.
AI Revolution in Cybersecurity
Both AI systems are engineered to uncover hidden vulnerabilities within digital infrastructures, boasting capabilities that can outpace human efforts in various cybersecurity tasks. The release of Mythos last April generated considerable buzz in the tech community, particularly after Anthropic reported discovering a security flaw in a legacy system that had remained undetected for nearly three decades.
Concerns about the ramifications of such powerful AI tools have escalated among finance ministers, central bankers, and industry leaders. Many are wary that models like Mythos could potentially jeopardize the integrity of financial systems.
Recently, the Governor of the Bank of England, Andrew Bailey, voiced frustration over UK banks' lack of access to Mythos, limiting their ability to fortify their digital systems.
OpenAI's Competitive Edge
Amidst this backdrop, OpenAI is ramping up its efforts. Alongside UK banking giants like Lloyds Banking Group, HSBC, and Nationwide—who will now benefit from GPT-5.5 Cyber—NatWest and Santander already have access through prior arrangements.
In an exclusive interview, former UK Chancellor George Osborne, now a senior executive at OpenAI, emphasized the necessity for these technologies to be managed responsibly. “These tools must be in the right hands,” he stated, advocating for their use in strengthening democratic institutions while preventing misuse.
[p][/p]
AI's Capabilities in Banking
Reports from the AI Security Institute indicate that both GPT-5.5 Cyber and Mythos demonstrate similar performance levels in cybersecurity assessments. Although Anthropic initially opened Mythos to a consortium of 42 companies—primarily tech firms in the US—OpenAI’s tool is now accessible to institutions in the EU, Japan, and Canada, among others.
The charge for usage by both companies reflects the investments made in their respective tools, with Anthropic directing $100 million towards Mythos previews.
Alan Woodward, a cybersecurity expert from Surrey University, highlighted the invaluable role AI can play in both modern and legacy banking systems. “AI can sift through millions of lines of code in minutes, identifying problems that even seasoned experts might overlook,” he explains, while also cautioning that human oversight remains essential due to the potential for false positives.
What's Next?
As the competition between OpenAI and Anthropic unfolds, the implications for UK banks and the broader financial sector will remain a heavily scrutinized topic. Stakeholders are keenly awaiting Anthropic's next moves and whether they will eventually offer access to Mythos for financial institutions.
Stay tuned for further updates as the landscape of AI in cybersecurity continues to evolve.
Source: BBC News - Technology
