The governments of the US, UK and Canada are raising concerns about the capabilities of this model.
Wall Street banks are beginning to test Anthropic Mythos’ new AI model as authorities raise concerns about cyber security around the world.
Anthropic opted against the public release of the Claude Mythos Preview, an AI league model that is better at generating revenue than the company’s previous launch.
Instead, Mythos was introduced to a select group of 40 “critical” technology and financial companies to test the model and strengthen their cyber defenses.
Alongside JP Morgan Chase – named by Anthropic as a partner – Bloomberg reported on April 10 that Goldman Sachs, Citigroup, Bank of America, and Morgan Stanley are also exploring the technology.
In its research, Anthropic noted that Mythos improved usage 181 times out of several hundred attempts, while Opus 4.6 had a success rate of almost 0pc.
“We didn’t explicitly train the Mythos preview to have these capabilities. Rather, they emerged as a byproduct of general code development, reasoning and autonomy,” it notes.
Other companies involved in Anthropic’s cybersecurity project called ‘Glasswing’ include Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, Linux Foundation, Microsoft, Nvidia and Palo Alto Networks.
Mythos has also raised concerns from government authorities, who are warning financial institutions about the risks posed by models like Mythos.
Last week, US Treasury Secretary Scott Bessent and Federal Reserve Chairman Jerome Powell warned Wall Street leaders to take Mythos “seriously”, while Canada’s top banks, as well as the parent company of the Toronto Stock Exchange and government departments, met on April 10 to discuss the cybersecurity risks suggested by the model.
Closer to home, the UK’s financial regulators are in talks with the National Cyber Security Centre, the Bank of England, the Financial Conduct Authority and HM Treasury to discuss the potential risks to their IT systems exposed by Mythos.
SiliconRepublic.com asked Bank of Ireland and AIB about whether they intend to use the model in their defense.
AI-powered cybersecurity attacks are becoming commonplace, with seemingly novice bad actors using the technology to easily compromise hundreds of firewalls.
Businesses, on the other hand, find themselves more exposed to cybersecurity threats.
Last year, Marks and Spencer, Co-op and Harrods, all famous names based in the UK, suffered from cyberattacks. Months later, cybercriminals stole the photos, names and addresses of about 8,000 kindergarten children in the country.
And later that year, US cybersecurity firm F5 suffered a major cyber attack that gave threat actors long-term access to some of its systems.
Don’t miss out on the information you need to succeed. Sign up for Daily BriefSilicon Republic’s digest of must-know sci-tech news.
