JPMorgan’s flirting with Anthropic’s AI toybox

JPMorgan meets the AI lab

JPMorgan is reportedly testing Anthropic’s AI tool Mythos, which is a very 2026 sentence to read over coffee. The bank isn’t just dabbling in “innovation theater” either — when the world’s biggest bank starts poking around an AI tool, the rest of Wall Street usually starts reaching for its own notebook.

Dimon: AI is the helper and the headache

Jamie Dimon also framed AI as a threat, which tracks with the vibe: the same technology that can speed up compliance, research, and customer support can also create new security, operational, and reputational risks. In other words, it’s the corporate version of giving a raccoon a credit card — exciting, but you’d better watch it closely.

Why investors should care

For JPM investors, this is less about one tool and more about strategy. Big banks have been racing to automate more work, cut costs, and make employees more productive, and AI could be a big piece of that puzzle.

But there’s a catch:

• If AI helps JPM run leaner, that’s a long-term margin tailwind.
• If it creates compliance or model-risk headaches, the cost of “moving fast” can get expensive, fast.

Big picture: JPMorgan testing Anthropic’s AI is another reminder that banks want the upside of AI without getting flattened by the downside. Easy in theory, messy in practice — which is basically the whole AI trade right now.