Anthropic to Brief Financial Stability Board on Mythos Cyber Findings
Beyond financial systems, Mythos is also exposing weaknesses in the security architectures relied upon by some of the world’s largest technology companies.
News
- Anthropic to Brief Financial Stability Board on Mythos Cyber Findings
- OpenAI Explores Legal Options Amid Tensions With Apple
- AI Has Outpaced How Companies Measure Developer Productivity, Report Finds
- AI Dispatch | May 8 - 14
- du Launches Sovereign Industrial AI Platform for UAE Manufacturers
- Huang Foundation Donates $108M CoreWeave Compute for AI Research
Released in preview over a month ago, Anthropic’s cybersecurity model, Mythos, continues to keep global leaders, particularly in financial systems, on their toes. It has been designed to detect decades-old vulnerabilities in web browsers, infrastructure, and software, and has emerged as a critical milestone in AI and cybersecurity advancements.
Recognizing the AI tools’ rapid evolution surpassing current regulatory frameworks, Andrew Bailey, the governor of the Bank of England and chair of the Financial Stability Board (FSB) of global regulators, last month said, “It is a very serious challenge for all of us. It reminds us how fast the AI world moves.”
Now, the AI startup is set to brief the FSB on cyber vulnerabilities in the global financial system identified by its latest AI model.
According to Financial Times, the Claude-maker will discuss the capabilities of its model with leading finance ministries and central banks from the FSB at Bailey’s request.
Cybersecurity experts have warned that the advanced system is being misused as a foundation for sophisticated cyberattacks, posing a risk to the banking industry reliant on legacy technology. Early on, Mythos was made available to select American firms, including JPMorgan and Bank of America, before being rolled out to British banks.
Beyond financial security, Mythos has left leading tech giants in a pickle regarding the safety measures employed. Cybersecurity startup Calif has claimed to breach Apple’s most advanced security systems in less than a week, prompting Mythos to accelerate the process.
Initiated on April 25, when a researcher discovered the first bugs, Calif identified vulnerabilities inside macOS running on Apple’s M5 hardware. In just a few days, the team combined two flaws into an exploit chain that escalated a normal user account to full root access on a MacBook running macOS 26.4.1. Reports suggest the researchers created what can be described as the first publicly known macOS kernel memory corruption exploit capable of bypassing Apple’s new Memory Integrity Enforcement, or MIE, system.
Introduced in 2015, MIE was defined as a major hardware-assisted defense mechanism designed to stop memory corruption attacks, which took five years to develop and billions of dollars in security investment. The exploit cost Calif $35,000 to build, while Apple poured billions in at an alarming rate, leaving many researchers shocked and concerned.
Notably, Calif has decided to hand-deliver its findings directly to Apple’s headquarters instead of submitting the vulnerability through normal channels.
