The AI cracks are starting to show at law firms
Law firms have poured vast sums into AI in a race to modernise, but ‘hallucinations‘ and security concerns are exposing the technology’s risks, writes Maria Ward-Brennan.
There has been intense pressure on legal, advisory, and consultancy firms to invest in, embed, and deploy AI over the last couple of years, but this pressure is now starting to crack the foundations of the models they’ve relied on for years.
City law firms have been ploughing hundreds of millions of pounds into AI, with a key focus on tech investment to stay ahead of what clients want and expect. The importance of tech spend has led private equity to surge as UK law firms seek external capital to fund expensive AI budgets.
It is a pivotal point in history for this industry, especially for lawyers, who – despite being traditionally risk-averse – are putting more and more of their chips on AI.
This funding has resulted in a booming legal AI startup industry, with Harvey reaching an $11bn valuation and Legora valued at over $5.5bn. There is so much money in this sector that these companies are trying to out-celebrity each other in market campaigns.
Harvey recruited the actor who plays Harvey Specter in the TV show Suits as its first-ever brand ambassador. But then Legora recently one-upped it by hiring actor Jude Law and running a ‘Law just got more attractive’ campaign that’s been all over London, including in CityAM.
Law firms are throwing cash – both from their profit pools and borrowed money – into AI, but it isn’t always smooth sailing.
‘That is not even a real law’
Top firms have been encouraging their juniors, and in some cases, paying, to ‘train’ up their AI model, but even with all the training, ‘AI hallucinations’ (otherwise known as ‘getting it wrong’) have exposed serious flaws, leading to costly mistakes.
This week, it was revealed that elite US law firm Sullivan & Cromwell had to apologise to a judge after its restructuring team made a filing in a high-profile case that contained multiple AI-generated ‘hallucinations’.
The head of its restructuring practice, Andrew Dietderich, apologised in a letter to the New York federal judge for mistakes that included misquoting the US bankruptcy code and citing cases incorrectly in a court filing.
The elite law firm, which typically charges $3,000 per hour, told the court that the firm maintained “rigorous” standards when using AI tools and that it “instructs lawyers to ‘trust nothing and verify everything’”.
But this isn’t the first time this has happened in the legal world. In the UK, courts have had to review cases after lawyers relied on citations and quotations generated by AI tools that proved entirely fictitious.
A senior High Court judge had to warn lawyers last year that the court has a range of powers, including referring matters to regulators, imposing wasted-cost orders, and imposing contempt or even criminal proceedings.
Look out, the fraudsters are coming
But the bad PR and financial hits from AI mistakes are not the only concern for firms.
Cybersecurity issues have been dominating the headlines following a warning over Anthropic’s Mythos, but law firms have already been feeling more anxious about cybersecurity.
In a fresh Law Society report, respondents said cybersecurity was the defining challenge for law firms.
Stewarts Law recently reported that criminals have been sending emails or faxes to the public pretending to be the law firm. One of these fake messages, sent as Stewarts, was sent to a member of the CityAM team, pretending to reach out about claims from representative clients in a fraud matter.
In a sector that holds a lot of client information and a large amount of money, it is already an ideal target for cybercriminals, who, ironically, are armed with AI advancements to facilitate their fraud.
Law firms will have to keep spinning the plates, even as they hope they don’t come crashing down.
Eyes on the Law is a weekly column by Maria Ward-Brennan focused on the legal sector.