By Priya Raman

Judges in England and Wales now have a cautious green light to use artificial intelligence, under a new position statement reported by The Law Society Gazette. The statement permits limited judicial engagement with AI tools. It stops well short of a general endorsement, and that gap matters.

The guidance arrives from the judicial leadership responsible for setting conduct standards for the bench in England and Wales. No case reference applies here; this is regulatory guidance, not a ruling. The date of publication accompanies the statement itself. Its subject is narrow: what judges may, and may not, do with generative AI systems in the course of their work.

What does the statement actually allow?

The headline promise is permission. Judges may use AI, cautiously. That is the entire concession. Absent from the report, at least so far, is any itemised list of approved tools, any approved-use register, or any technical standard against which a court's AI output might be audited.

Practitioners should note what the statement does not say. It does not say AI-generated material is admissible by virtue of judicial blessing. It does not say judges may delegate reasoning to a model. And it does not, on the reporting available, resolve who bears responsibility if a judge relies on a flawed AI output in a judgment. Every promise of efficiency here carries a caveat.

Why did the judiciary act now?

Courts across jurisdictions have spent two years absorbing guidance on AI, mostly aimed at lawyers and litigants. Judges themselves sat in a grey zone: using the tools, or declining them, without formal cover. This position statement closes that gap for the England and Wales bench. It follows a pattern already visible in other common-law jurisdictions, where senior courts have issued AI guidance for judicial officers.

The move is cautious by design. Judicial reasoning carries authority that no probabilistic text generator can warrant. A model can fabricate; a judge cannot. The statement's restraint, therefore, reads less as technophobia and more as an acknowledgement of that asymmetry.

What does it change for practitioners?

For solicitors and barristers, the practical consequence is indirect but real. If judges may consult AI tools, the courtroom is no longer an AI-free zone by assumption. That shifts the burden onto counsel: to check submissions for AI-generated error, to disclose AI involvement where rules require it, and to anticipate bench questions about the provenance of assembled material.

The statement also signals a drafting environment in transition. Guidance for the judiciary tends to anchor later guidance for the profession. Firms should expect follow-on rules, and should treat this statement as an early marker rather than a finished framework.

Takeaway: judges may now use AI in a limited way, but the statement grants permission without granting trust, and the unanswered questions outnumber the settled ones.