Why AI Legal Assistants Wont Fix the Broken British Court System

The British justice system is drowning. Right now, the number of cases waiting to be heard in the Crown Courts of England and Wales has soared past 80,000. That is more than double the pre-pandemic backlog from 2019. Things are so bad that thousands of trials are being scheduled for 2028, and a few nightmare cases are pushed all the way out to 2030. Victims of crime are waiting years just to see a courtroom.

So, what is the government's grand plan? Algorithms.

Deputy Prime Minister David Lammy is using London Tech Week to announce a major push for virtual legal assistants powered by artificial intelligence. The Ministry of Justice wants to trial these tools in Crown Courts to handle routine casework, analyze legal arguments, and assist judges with case scheduling. On paper, it sounds brilliant. In reality, it looks like an attempt to paper over the cracks of a crumbling infrastructure with shiny new software.

Legal professionals aren't buying the hype. While technology can help clear an inbox, it cannot replace the physical courts, human lawyers, and support staff required to keep the wheels of justice turning.

The Illusion of the Automated Courtroom

The government’s new strategy focuses heavily on driving productivity through automation. Under the plan, AI legal assistants will be built alongside legal experts and developers to help lawyers with research and case analysis. Judges will get a separate tool designed to spot trial-ready cases and group similar hearings together.

The administration already points to success within the Probation Service, where every officer has been handed a tool called Justice Transcribe. The Ministry of Justice says this transcription app has already logged over 800,000 meetings, saving roughly 133,000 hours of manual data entry.

That is fine for internal admin. But a court of law is a completely different beast.

When you introduce generative technology into criminal trials, you aren't just summarizing a meeting; you are dealing with liberty, evidence, and public safety. The legal profession is already littered with examples of what happens when software goes wrong. Just last year, during an £89 million damages case involving the Qatar National Bank, a claimant submitted 45 case law citations. It turned out 18 of them were entirely made up by an AI tool. The solicitor had to admit they just trusted the machine's output without checking.

Even law enforcement has fallen into the trap. A recent investigation into why Maccabi Tel Aviv football fans were banned from a match against Aston Villa revealed that West Midlands Police used a report generated by Microsoft Copilot. The AI suffered a hallucination and invented a completely non-existent match between Tel Aviv and West Ham to justify the security risk.

If police forces and corporate lawyers are already falling victim to these glitches, forcing this technology into overworked Crown Courts is a massive gamble.

Why Software Can’t Pay the Bills

Ian Jeffery, the chief executive of the Law Society of England and Wales, didn't mince words when responding to the government's announcement. He made it clear that while new tech might enhance access to justice, it cannot replace vital funding and additional court staff.

That is the core issue. You can have the fastest document-parsing algorithm in the world, but it won't matter if there isn't a physical courtroom available, a judge to sit on the bench, or a barrister to argue the case.

The backlog isn't happening because lawyers are slow typists. It is happening because of years of systemic underfunding.

• Legal aid rates have been depressed for decades, driving criminal barristers out of the profession entirely.
• Court buildings are literally falling apart, with broken heating systems and leaking roofs regularly forcing daily closures.
• Staffing shortages mean cases get dropped at the last minute because there aren't enough clerks or security officers to run the room.

An AI legal assistant can't fix a broken boiler. It can't step in as a defense solicitor when legal aid funding fails to cover the cost of a trial. By framing tech as the primary solution to a record-breaking backlog, the government risks dodging the real, expensive work of rebuilding the justice system from the ground up.

The Dangerous Road Toward Scaled-Back Trials

Perhaps the most worrying aspect of this sudden push for digital efficiency is what follows in its wake. Alongside the tech announcement, the government has floated plans to scale back certain jury trials to help clear the logjam.

This reveals the underlying danger of the efficiency-first mindset. When a public service is starved of cash, technology isn't used to empower human workers; it's used to justify cutting them out of the loop. If the goal shifts from delivering fair justice to simply moving numbers through a dashboard, the integrity of the whole legal system is compromised.

The Law Society is demanding that the outcomes of these court pilots be made entirely public before any wider rollout. That is a bare minimum requirement. We need total transparency on how these tools behave, how often they hallucinate, and whether they introduce bias into the listing of criminal trials.

What Needs to Happen Instead

Technology has a place in the modern office, and no one is arguing that courts should stick to paper ledgers and fax machines. But if the Ministry of Justice wants to actually fix the backlog rather than just manage the optics, the approach has to change.

First, the court pilots must be strictly evaluated by independent oversight bodies, not just the tech companies building them. If an algorithm suggests grouping cases or flags a file as "trial-ready," there needs to be a clear, auditable trail explaining why.

Second, the government needs to match every pound spent on digital infrastructure with a direct investment in human capital. This means raising legal aid fees to attract new talent back to criminal law, hiring more court clerks, and repairing the physical buildings that are currently unusable.

If you want to clear a backlog of 80,000 cases, you need more people power, not just more processing power. Software should be a tool for the staff we have, not an excuse to avoid hiring the staff we need.