Few industries appear to have more potential for disruption by artificial intelligence than the law. Like games such as Go, which DeepMind took on to demonstrate the power of neural networks, legal systems have sets of rules and precedents. Give an AI model enough data and it can pass its bar examination.

So the law has become the target of many AI start-ups, such as Harvey, a US venture that in July raised $100mn at a $1.5bn valuation from investors including Google Ventures and OpenAI. There has been a parallel rush of activity in the UK, with three legal AI start-ups — Genie AI, Luminance and Robin AI — raising more than $100mn in combined funding last year.

The UK’s growth is a product of the overlay of two successful industries, one old and one new: the commercial law and the cluster of AI start-ups, particularly in London and Cambridge. “We have a rich heritage and some incredible talent. It is a great competitive advantage,” says Eleanor Lightbody, chief executive of Luminance, which was founded in 2015 in Cambridge.

Slaughter and May was among Luminance’s early investors, along with the late Mike Lynch’s Invoke Capital, and it initially focused on reforming how corporate law firms operate. They provide expensive due diligence for transactions such as mergers and acquisitions that involve junior lawyers sifting through thousands of documents. That sounds like a plum target for disruption.

But things have not quite worked out like that, so far. Much of the recent expansion in legal AI has come not from integrating the technology into law firms but automating the drafting of contracts by companies: Genie AI’s users include many smaller manufacturing and building companies. More than changing the legal industry, it is allowing corporations to do more legal work themselves.

Although many law firms have experimented with AI, the tradition of billing by the hour of human labour means that using it often equates to reducing revenues. “My recommendation to big law firms is not to bother too much with generative AI. They are being paid for their human expertise, not speed,” says Rafie Faruq, Genie AI’s co-founder and chief executive.  

The market for contract writing in other companies is also far larger. Legal departments are often flooded with internal demands to sign off deals with clients and suppliers. “There are people on every floor in every office in the world receiving contacts, reviewing them and making decisions,” says Lightbody, who joined Luminance from Darktrace, another Lynch-backed AI venture.

Companies can upload thousands of existing contracts into AI models, which analyse the clauses and terms they tend to use. The technology can use that knowledge to help to draft new ones, showing where changes are needed. Luminance’s platform highlights in red and amber clauses that require human attention and can suggest reworkings to ensure they comply.

This is rather like having an experienced lawyer by the side of an employee who is preparing a contract: they can easily and cheaply call on a large database of knowledge. The ability to search past contracts efficiently has other benefits. Companies can, for example, easily find which clients they must inform of a cyber attack instead of hiring a law firm to carry out a review.

Much of the work is routine, and Faruq says in-house lawyers are happy to be relieved of it: “They don’t want to write boring contracts. They want to work on higher-value deals.” But companies also rely on their lawyers for institutional knowledge. An AI model that reads every contract is a potential rival.

The legal AI industry faces its own uncertainties. A lot of capital has flowed into start-ups that employ standard AI models, customised with their own technology. (Robin AI is based on Anthropic while Genie AI uses Anthropic and OpenAI). If the underlying technology is similar and the training data of legal statutes and precedents is identical, the barriers to entry may prove low.

Luminance was founded by machine-learning experts and has its own AI model, trained on 150mn legal documents. Genie puts more emphasis on its library of 1,000 templates that an AI agent draws on to write legal contracts from scratch. “It is impossible to beat the big providers on AI,” Faruq says.

The harsh reality that not everyone will survive is coming to the legal AI industry in the future. For now, the flow of capital indicates not only the potential of this technology outside the law, but the UK’s competitive strength.