Ayinde v Haringey explained: AI, fake citations and your duty

The case that put AI's biggest risk in front of the High Court — and the clear duty the court set out in response. Required reading for any solicitor using AI.

If there is one case every solicitor using AI should know, it is R (Ayinde) v London Borough of Haringey, heard together with Al-Haroun v Qatar National Bank — [2025] EWHC 1383 (Admin). It is the moment the profession's biggest AI risk landed in front of the High Court, and the court's response set the standard.

What happened

In short: submissions placed before the court referenced authorities that did not exist — non-existent cases of exactly the kind a large language model will fabricate when asked for supporting law. When the fictitious citations were questioned, they could not be justified. This is the characteristic AI failure mode: fluent, confident, and entirely invented.

What the court said

Dame Victoria Sharp P, giving judgment, was direct. At paragraphs 7 and 8 she set out the duty in terms worth holding onto:

Those who use artificial intelligence to conduct legal research notwithstanding these risks have a professional duty to check the accuracy of such research by reference to authoritative sources, before using it in the course of their professional work.

And critically, she located that duty within the ordinary architecture of professional responsibility:

This duty rests on lawyers who use artificial intelligence to conduct research themselves or rely on the work of others who have done so. This is no different from the responsibility of a lawyer who relies on the work of a trainee solicitor or a pupil barrister, for example, or on information obtained from an internet search.

Why it matters so much

The ruling does two things. First, it confirms that the duty to verify is not new or special — it is your existing duty of supervision and competence, applied to a new source of error. Second, it makes clear there is no hiding behind the tool: using AI is a choice, and with it comes the responsibility to check what the AI produces. The court was not telling lawyers to stop using AI. It was telling them to use it properly.

What it means for your practice

The practical upshot is a single, non-negotiable rule: never rely on AI-generated authority you have not verified against the original source. Not a case, not a statute, not a "market standard," not a quotation. If the model asserts it and you can't check it, you don't use it.

This is precisely the discipline the skills in The Lawyer's Prompt build in. They never invent authority, and they flag anything unverifiable with [VERIFY] so it cannot quietly slip into your work. The standard the court set in Ayinde isn't a constraint bolted onto the tools — it's the thing they're designed around.

Read next: how to verify AI output and AI and SRA compliance.