AI in the justice system

For years, much of the legal profession’s conversation around artificial intelligence has been dominated by fear: hallucinated cases, confidentiality risks, regulatory breaches, and the now-familiar warnings about lawyers outsourcing judgment to machines.

Just recently we all heard about the Sullivan’s AI mishap (or more than this?).

Those concerns are real.
But they are only half the story.

Recent comments from Sir Colin Birss, Chancellor of the High Court, offer a more sophisticated, and arguably more useful—view: AI is not inherently a threat to legal work. Used properly, securely, and critically, it may materially improve it. This is exactly what I emphasised in the recent post. From helping judges anonymise sensitive judgments to identifying internal inconsistencies in draft decisions, the judiciary is increasingly treating AI as an advanced professional support tool.

Sir Colin’s position is notably pragmatic. Judges in England and Wales are not prohibited from using AI. Instead, the framework is clear: whilst the judge retains sole responsibility it must ensure that outputs are verified and systems are secured. AI is not here to replace decision making, and should be treated as assistant.

I totally agree that AI is proving to be useful in (amongst others):

1. Better anonymisation and reduced “jigsaw identification” risk

AI can identify combinations of seemingly minor details that could reveal identities even when names are removed, which is a critical issue in family and sensitive matters.

2. Internal consistency checking

Sir Colin himself has described using secure Copilot tools to flag contradictions or unclear drafting in judgments.

NB: arbitration professionals rightly note that it is okay to verify what’s been drafted but not to use AI to replace your drafting, even of a factual matrix.

3. Administrative efficiency

Search, transcription, and information retrieval functions can reduce time spent on low-value process tasks and redirect focus to legal reasoning.

4. Access to justice

AI-assisted drafting may help litigants in person present claims more coherently than they otherwise could.

IMO, whilst access to justice is clearly easier, litigants in person are sadly too dependent on even ridiculous outcomes simply due to a lack of knowledge or experience. It may lead to wasted time, adversary costs, or emotional distress.

The real danger is not AI but abdication of professional judgment. It looks like the legal profession’s mistake is to confuse, at times, assistance with substitution.

AI cannot exercise legal judgment, assess strategic nuance or understand professional ethics. That is where competence remains entirely human.

You can access Sir Colin Birss’ full speech here: https://www.judiciary.uk/speech-by-the-chancellor-of-the-high-court-legal-professional-privilege-in-the-age-of-ai/