Most of the debate about AI and law falls into three camps. Here is what each believes, and the evidence behind it.
It's now my third month running The Redline, and I want to try something a little different. Usually I write about updates in the LegalTech world. This time, I want to share my own understanding of where AI and law meet, based on everything I've been reading across LegalTech newsletters and accounts.
As I've followed the conversation, I've noticed that most people fall into one of three groups. Each group is trying to answer the same question: what is the future of law? And, following from that, will AI mean fewer jobs in law?
As always, this article won't take a side. It will lay out what each group believes, and why.
The first group believes law is about much more than reviewing contracts. It's about judgement and relationships, and in their view, AI can't replace that.
Law often calls for objectivity, but people have an ability to pick up on nuance and understand how other people think and behave. This group argues that no amount of coding will ever let AI truly understand each individual person and the circumstances of each case. There's some research behind this. When researchers tested whether leading AI models could reason like real jurors in a mock trial, the models could not reliably reproduce how human jurors reach their decisions.1 Lord Neuberger, former President of the UK Supreme Court, made a related point in October 2026: people have a basic right to understand why a decision went against them, especially where their liberty, home or children are at stake.2
Interestingly, this group doesn't deny that AI will have a huge impact. Many of them accept that much of the contract work and admin in law will be taken over by AI. But they see judgement as so central to the job that, even if AI became 100% accurate and stopped hallucinating altogether, there would still be a role for lawyers. That role would simply shift towards the parts of the job that need judgement.
The second group believes people overestimate how much AI will change law.
They point out that AI is still far from completely accurate, and in law, a single mistake can be serious. On an independent legal research benchmark, the base model behind OpenAI's new legal tool answered only about two in five questions fully correctly.3 And the mistakes are already reaching the courts. In one Mississippi case, four lawyers were sanctioned after filings containing AI-hallucinated citations, including two who never used AI but signed off without checking.4 The SRA has been clear that using AI is never an excuse for getting something wrong.5
This group also argues that AI hasn't delivered the efficiency gains many expected. In a 2026 Bloomberg Law survey of more than 700 lawyers, 83% reported a high level of AI adoption, yet the analysis concluded AI was “not an efficiency bombshell.”6 And in the first half of 2026, demand for lawyers' billable hours actually rose 4.2%, more than double the usual rate.7
For this group, AI's role in law is clear: at best, it's an assistant. Useful, but always needing a lawyer to check its work, and certainly not about to replace the solicitors who draft and review contracts.
The third group believes AI is already changing how law works, and that this will reshape the profession.
They argue that AI is making many tasks faster and easier, especially the routine work that junior lawyers have traditionally done. As early as 2023, a Goldman Sachs report estimated that 44% of legal tasks could be automated by AI.8 Lord Neuberger has warned that AI could cause “a real fall-off in recruitment of young lawyers”, since AI-drafted research and advice is often as good as a newly qualified lawyer's, and produced more quickly and cheaply.2
If work takes less time, the billable hour becomes harder to justify. In September 2026, Citi, Morgan Stanley and Goldman Sachs began pressing their law firms to pass on AI savings.9 Meanwhile, new firms like Falcon have dropped hourly billing altogether, charging a flat monthly fee instead.10
For this group, the question isn't whether law will change, but how quickly.
It's important to remember that most lawyers don't sit neatly in one group. Many move between them, depending on the firm they work at, how much that firm encourages AI, and the training they've had themselves.
Take a firm that's very keen on using AI to draft contracts. Lawyers there are likely to sit in the first group, or even the third. Now take a firm that's sceptical about AI, worried about hallucinations and mistakes. Lawyers there may well sit in the second.
There's no right answer, and to be honest, what I believe doesn't really matter. What matters is understanding all three views, and why lawyers hold them.