Anthropic's Mythos AI tool could spark 'systemic financial market disruption,' a trade group warns

• An investor group says Anthropic's new AI tool, Mythos, could destabilize the entire financial system.
• The tool identifies code vulnerabilities, and could lead to a widespread data breach, the group said.
• Consequences could range from mass identity theft to "systemic financial market disruption," it said.

Anthropic's newest AI tool could spark chaos in markets, but not for the reasons investors have become accustomed to lately.

The AI giant — which has rolled out new tools and updates this year that have upended parts of the market — recently developed a new model aimed at improving cybersecurity. But the model risks exploiting a key vulnerability in the financial sector, introducing risks ranging from widespread identity theft to the destabilization of the financial system, the American Securities Association said.

In a public letter to the Treasury Secretary on Thursday, the trade group flagged concerns about Claude Mythos, the "general-purpose" AI model Anthropic announced in early April.

The model—which falls under Project Glasswing, the company's broader cybersecurity initiative—is able to locate "thousands of high-severity vulnerabilities" in code across "every major operating system and web browser," the company said on its website.

If used by bad actors, the tool could be used to hack into the Securities and Exchange Commission's Consolidated Audit Trail, a centralized database that contains investors' private information, the ASA speculated.

The group's letter comes about a week after reports said that Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened major US bank CEOs for an urgent meeting to flag the cyber risks posed by Mythos.

"The subject matter of this meeting confirms what ASA has warned about for years: the US Securities and Exchange Commission's (SEC) Consolidated Audit Trail (CAT) is a significant cybersecurity vulnerability waiting to be exploited. This is no longer a hypothetical. The threat is here, it is identified, and it has a name," the letter said, referring to Mythos.

The ASA has long opposed the use of the Consolidated Audit Trail, citing data privacy concerns.

The group outlined six specific risks it believes Mythos could pose to investors.

1. Mass identity theft. The personal info of retail investors could be exposed "in a single breach," the letter said.
2. Trading strategy and portfolio exposure. The breach could also expose sensitive information about investors' positions.
3. Explotation of financial market intelligence. The data could be used as an "extraordinary weapon" by foreign countries, like China or Russia.
4. Exploitation of software vulnerabilities. "Mythos excels precisely at finding decades-old, dormant flaws of the kind that permeate the middleware, data feeds, browsers, and operating systems that underpin CAT's vulnerable architecture," the group said.
5. Insider threat risk. The data could be used by "malicious insiders" at companies, the letter said.
6. "Systemic financial market disruption." A bad actor gaining access to the info could trigger mass liquidations of risk assets, leading to a cascade of "failures across the financial system," the group speculated.

The group outlined actions regulators could take, including suspending CAT and getting rid of the platform's collected data.

Anthropic, which described Mythos as a work in progress on its website, also noted the potential consequences of the technology if it were used by bad actors.

"Given the rate of AI progress, it will not be long before such capabilities proliferate, potentially beyond actors who are committed to deploying them safely. The fallout—for economies, public safety, and national security—could be severe," it said on its website.

Anthropic and the US Treasury did not immediately respond to a request for comment from Business Insider.