'The market's in seek and destroy mode': The new AI model scaring lawyers and legal firms

Anthropic, one of the biggest and most influential tech companies in the world, is launching a new model: Claude Opus 4.6.

Until now, this would mostly be big news for techies, where Anthropic is admired as the maker of Claude Code, the code-writing AI tool which many engineers say is taking over their work entirely. All of a sudden, however, the impact of these tools is being felt more widely, after a seemingly small release from Anthropic shook some sections of the stock market.

Earlier this week, Anthropic released a plug-in to their Claude chatbot, which added new tools for legal analysis. This relatively minor update sent shares of several large legal data firms tumbling.

Thomson Reuters, which owns legal database Westlaw, fell nearly 16%. Analytics company RELX fell 12%.

This may seem strange to outsiders because Anthropic is still relatively unknown outside the world of tech. A poll by US research firm Blue Rose Research at the end of last year found that it was known by less than 5% of the population.

But analysts say that market participants, impressed by the advances in coding, are starting to wonder if the same gains can be made in other sectors. "The market's in seek and destroy mode," says James Sym, partner at London-based equity firm Goodhart.

"It's just trying to find the next loser from AI. That's what you've seen in the last couple of days." Anthropic's latest release may add to this sense of urgency, because it is directly aimed at knowledge workers.

As well as a number of improvements to coding - the ability to handle large codebases, for instance, and longer tasks - the new model is designed to deal with the kind of problems non-coders face in apps such as Excel and PowerPoint, where Claude will now be able to work directly. "Users can build slides from a corporate template, restructure a storyline, convert bullets into diagrams, or generate a full deck from a description - all without leaving the app," the firm says, although users will have to pay for the privilege.

It adds that this is "our most capable model for all enterprise and knowledge work". Anthropic says its new model outperforms its old model on a number of key benchmarks, pointing to an assessment by Norway's Sovereign Investment, which found that "across 40 cybersecurity investigations, Claude Opus 4.6 produced the best results 38 of 40 times in blind ranking against Claude 4.5 models".

Read more from Sky News:Man who stabbed nine-year-old in the heart guilty of murderEven Starmer's allies are unsure about his future But the firm is playing down the impact of its knowledge work tools, directing people to statements made by legal software-makers who build specialist tools using Claude Code. "There is an important difference between a plugin and operating a collaborative, matter-centric, production-grade platform used by hundreds of the world's leading legal teams," says Max Junestrand, CEO at Legora, an AI tool for lawyers.

The market is not seeing things in such a nuanced way, says Sym of Goodhart, who suggests this could be the "canary in the coal mine" for the end of the exuberance around AI. "If you think about how bubbles in history have evolved over time, they normally follow a bit of a pattern, and what happens is fewer and fewer companies seem to be the winners.

That is what's obviously happened at the moment. "This may well be just part of that normal pattern where you're seeing the market decide, actually, it is only going to be a very narrow set of people who win.

And of course, the step after that is the whole bubble bursts. And that is what we could be looking at.".