Has JPMorgan actually laid off workers because of AI?: Yes — CEO Jamie Dimon confirmed in February 2026 that JPMorgan has 'displaced people from AI', while offering internal redeployment. The bank committed $600 million annually to retraining.Source: CNBC
What is Project Glasswing and why does JPMorgan have access?: Project Glasswing is Anthropic's restricted-access programme giving select partners pre-deployment access to Claude Mythos. JPMorgan is one of twelve original members, giving it frontier AI capabilities ahead of public release.Source: Anthropic
What is JPMorgan spending on AI retraining?: $600 million per year, tied to performance reviews for 65,000 engineers. MetaIntro found only 6% of companies globally are reskilling for AI — making JPMorgan's programme a significant outlier.Source: CNBC / MetaIntro
Why did the Fed and Treasury meet about AI and banks in April 2026?: The Bessent-Powell emergency meeting of 8 April was convened over cybersecurity risks from AI models including Mythos — the same model JPMorgan has privileged access to through Project Glasswing.Source: UK AI Security Institute / Axios

Background

JPMorgan Chase is the largest US bank by assets, with operations spanning investment banking, consumer finance, commercial banking, and asset management. In February 2026, CEO Jamie Dimon confirmed at the bank's investor meeting — as reported by CNBC — that JPMorgan has "displaced people from AI" and offers them internal redeployment. The bank committed $600 million annually to retraining and tied engineer performance reviews to AI tool adoption across 65,000 staff. JPMorgan is one of the twelve original Glasswing partners with privileged pre-deployment access to Anthropic's Claude Mythos Preview.

Dimon's public acknowledgement is significant because it came from an institution with every incentive to downplay internal disruption. MetaIntro research circulating at the same time found only 6% of companies globally are actually reskilling workers for AI — making JPMorgan's retraining programme a statistical outlier by a factor of roughly 16. The bank's move to tie engineer performance reviews to AI adoption is an early example of formal AI-productivity requirements embedded in employment contracts at a global systemically important financial institution.

JPMorgan sits at the intersection of three threads in this beat: it displaces workers (admitted by its CEO), it retrains workers (a near-Singular commitment), and it has access to the model whose cybersecurity capability prompted the Bessent-Powell emergency meeting of 8 April. Its Glasswing access means JPMorgan's own AI deployment is already operating at or near the frontier capability the UK's AISI confirmed — the 20-hour autonomous task completion standard — while the rest of the industry absorbs Anthropic's shift from ASL thresholds to autonomy-focused risk models.

How the World Sees Them

Labour

The $600M retraining pledge is a benchmark no other bank has matched; critics note it is still a fraction of the displacement it is causing.

UK regulators

The Bank of England's FPC flagged agentic AI in financial services as a rapidly increasing systemic risk in April 2026 — JPMorgan's Glasswing access is the sharpest instance of this concern.

US finance sector

JPMorgan is the only major bank publicly acknowledging AI-driven internal displacement; competitors are watching how its Glasswing access and retraining commitment land with regulators.