Story | Folio

AI & Law

Freshfields and Anthropic announce joint development of AI legal tools as the Bank of England and UK Finance push banks to deploy AI cyber defences against frontier model threats

Freshfields and Anthropic have announced a collaboration to jointly develop AI legal tools, according to Reuters, in one of the most high-profile AI partnerships between a Magic Circle firm and a frontier AI developer. The deal places Freshfields at the leading edge of City firm AI strategy, and the nature of the joint development arrangement — rather than a simple licencing relationship — suggests the firm is seeking to shape how large language models are deployed in legal workflows, rather than simply consuming off-the-shelf products. No financial terms or specific use cases were disclosed in available sources. Separately, the Bank of England and UK Finance (the trade body representing UK banks and financial services firms) have jointly warned banks and insurers to strengthen their cyber defences using AI, citing the threat posed by emerging frontier AI models such as Anthropic's Mythos. The BoE guidance links cyber resilience directly to the capabilities of the same class of AI systems that Freshfields is now partnering to deploy, creating a sharp regulatory context for City firms considering their own AI adoption: the tools being built are simultaneously the subject of systemic risk warnings from the UK's central bank. The Lloyd's Market Association (LMA) has also published a new AI adoption toolkit this week, building on its April 2026 survey on AI risk management, to guide managing agents through governance frameworks as AI adoption accelerates across the Lloyd's market.

Why this matters

The Freshfields-Anthropic partnership signals a structural shift in how elite law firms are positioning themselves relative to AI developers — moving from technology consumers to co-developers with the ability to influence model training, deployment parameters, and liability allocation in the tools they use. This creates immediate demand for AI governance policy work, technology licence and IP assignment review, and data processing agreement drafting, as any joint development arrangement involving client data will raise questions under UK GDPR and the EU AI Act. The BoE's simultaneous warning about frontier model cyber risks — citing Anthropic's Mythos specifically — adds a regulatory dimension: firms deploying AI tools built on the same model architectures flagged as systemic risks will need to demonstrate to regulators that their governance frameworks are adequate. The LMA toolkit launch reflects the same pressure hitting the insurance sector.

On the Ground

A trainee on an AI governance or technology licensing matter would be reviewing and marking up data processing agreements between the law firm and the AI developer, drafting AI governance policy sections for the firm's internal compliance framework, and preparing regulatory impact assessment memos summarising the EU AI Act's requirements for high-risk AI system deployment in professional services contexts.

Interview prep

Soundbite

Freshfields co-developing AI tools with Anthropic shifts the firm from consumer to co-architect — and raises direct UK GDPR and EU AI Act exposure.

Question you might get

“If Freshfields is jointly developing AI tools with Anthropic that will process client documents, what are the key legal risks the firm needs to manage, and which regulatory frameworks apply?”

Full answer

Freshfields and Anthropic have announced a joint AI legal tools development partnership — a meaningful step beyond the standard licencing deals most firms have struck with technology vendors. By co-developing tools, Freshfields takes on influence over model behaviour but also greater exposure: questions about IP ownership, data processing obligations under UK GDPR, and liability for outputs all become live in a way they are not for simple users of off-the-shelf products. The Bank of England's simultaneous warning about Anthropic's frontier models as cyber threats adds a striking regulatory context — the firm is partnering with a company whose most advanced model the UK's central bank is publicly flagging as a systemic risk. The broader trend is clear: elite firms that fail to invest in proprietary AI capability will face competitive disadvantage in both speed and cost, forcing a wave of similar partnerships across the market. The regulatory governance question is the one I'd want to dig into most closely in any firm interview.

Sources

My notes

saved

Discussion

Loading…