What happened
Warren, who chairs the Senate Banking Subcommittee on Economic Policy, filed the bill in Washington on Wednesday morning, according to CryptoBriefing's report timestamped 10:19 UTC. The text would require any financial institution under federal supervision to file periodic disclosures detailing where AI systems are used in trading, credit decisioning, compliance, and risk management.
The filings would go to the institution's primary regulator, with a summary published by the Financial Stability Oversight Council. That last detail matters. It pushes AI from an internal compliance question into a public, cross-agency one, with FSOC sitting on the same data it uses to flag emerging systemic threats.
The bill does not mandate any specific limit on AI use. It mandates that the use be visible.
Why it matters
Crypto trading infrastructure has quietly become one of the more AI-saturated corners of finance. Market makers like Wintermute and Cumberland DRW have publicly described machine-learning execution stacks, and several US-regulated prime brokers route flow through models they have not been required to itemize. If Warren's bill becomes law, those models stop being a black box for regulators.
The Block reported earlier this year that more than 60% of spot BTC liquidity on Coinbase during US hours is touched by algorithmic systems, many AI-trained. A disclosure regime would not slow that flow. It would price it.
Counterparties, custodians, and ETF authorized participants would start asking for the filings before extending credit lines, the same way they ask for VaR reports today. That changes who gets capital and at what cost.
