The growing use of advanced AI tools is starting to unsettle parts of the US financial ecosystem, and the latest warning centres around how such technology could interact with one of the market’s most sensitive data systems. Anthropic’s Mythos AI model has triggered fresh concerns after the American Securities Association (ASA) flagged the possibility of misuse involving the market-tracking infrastructure overseen by the US Securities and Exchange Commission. The group believes the combination of large-scale financial data and increasingly capable AI systems could create a new kind of vulnerability that regulators may not be fully prepared for.
Concerns grow around misuse of trading database
At the heart of the issue is the Consolidated Audit Trail (CAT), a system built to give regulators a detailed view of trading activity across US markets. While the database has been useful for surveillance and investigations, the ASA argues that its design does not account for the kind of AI-driven threats that are now emerging.
In its communication to US Treasury Secretary Scott Bessent, ASA President Chris Iacovella warned that tools like Mythos could make it easier for malicious actors to tap into the system. According to the group, this could lead to large-scale identity theft, exposure of individual portfolios, and a rise in insider risks if such data is analysed or reconstructed using AI.
The association has pushed for immediate safeguards, including a halt on collecting personal data from retail investors and even the deletion of existing records stored in the CAT. This marks one of the strongest calls yet to rethink how much information the system should hold.
Experts say the risk is not limited to personal identifiers. Sultan Meghji pointed out that even anonymised datasets can carry strategic value when processed with modern AI tools. He noted that the cost and effort required to extract meaningful insights from such data have dropped sharply.
“The concern isn’t just personal identity data,” Meghji said, adding that what earlier required highly specialised teams can now be done with far fewer resources. “Mythos and systems like it push that further, enabling synthesis and misuse of decades of market data at industrial scale,” he said.
The issue has also caught the attention of top policymakers. Recent discussions involving Jerome Powell and leading Wall Street executives reportedly included a review of Mythos and its possible impact. The ASA has used this to underline its argument that the CAT could become a major cybersecurity weak point if not updated.
The regulatory debate around the system is already evolving. The SEC, led by Chairman Paul Atkins, has invited public feedback on potential changes, suggesting that even foundational aspects of the CAT could be reconsidered. This comes as authorities balance the need for market transparency with rising concerns around data security in the age of AI.
