Eliot Raman Jones

Mitsubishi UFJ Financial Group (MUFG) has designed a proprietary version of Anthropic’s Claude Code AI harness because the coding tool was too effective at accessing information and bypassing the bank’s security and compliance guidelines. Stu Taylor, the bank’s new head of global markets AI for all regions except Japan, tells WatersTechnology that while the AI tools the bank utilizes are all human-in-the-loop, public agentic AI “harnesses”—tools like Claude Code, Cursor, and GitHub Copilot

This is my fourth British summer heatwave I’ve experienced while at WatersTechnology, and my 28th altogether. While each summer is hotter than the last one, there are a few constants each year. The first is that British architecture’s lack of preparedness for intense heat means that we can’t deal with temperatures that some people in other countries would consider normal, which leads to jeering on social media.

At JP Morgan’s annual general meeting on May 19, CEO Jamie Dimon was asked about the bank’s primary risk exposures and whether the growing use of artificial intelligence and digital tools meant that operational risk had become the bank’s biggest threat. “I think the largest threat to the company is cyber. We’ve written about that before,” Dimon said. “If you consider cyber [-security risk] operational [risk], the answer would be yes.” Dimon’s hesitation over whether to classify cyber as