The Promise and the Perils of Artificial Intelligence in Court Work

The Promise and the Perils of Artificial Intelligence in Court Work

Artificial Intelligence flows through the hallowed halls of justice as the morning mist—pervasive, transformative, and unstoppable. Courts and lawyers alike should embrace AI — with wisdom, not fear.
Frost reminds us, “The best way out is always through.”
No machine can ever match the human soul’s eternal quest for justice.
That fire burns beyond all programming.

by GK Ganesan May 27, 2025 8228 Views

[1]. The AI Revolution in Our Courtrooms: A Wake-Up Call.

Artificial intelligence will fundamentally transform how judges and court-going lawyers work. It would be ridiculous to think otherwise. The facts prove it.

The statistics on what is happening right now, in May of 2025, are staggering, and verifiable:

According to a survey conducted by Thomson Reuters’: 26% of legal professionals are now actively using generative AI, nearly doubling from 14%– that was just one year ago. This is not speculation – it is happening in real time.¹

[2]. Three transformations demand your immediate attention.

One, AI has revolutionised legal research and document analysis.

Implementing AI has slashed document review costs (fees paid to lawyers to examine case materials) by 60 to 80%.²

Bloomberg Law reports that its AI contract management platform has reduced contract drafting times of Fortune 500 companies: from a soporific 6 weeks to just 72 hours.³

During the Meta Platforms litigation (a case about inflated ad audience claims), AI identified 14 non-compliant ‘data usage’ clauses (these are contractual terms violating data protection laws) across 12,000 ‘legacy contracts’: i.e., old agreements still in effect today.⁴

Two, AI gives us unprecedented strategic insights.

The Huckabee v Bloomberg litigation shows this. It was a copyright lawsuit over AI training.

Defence teams utilised ‘predictive analytics’ tools. These tools were trained for 15 years on what were known as ‘First Amendment rulings’. These were court decisions protecting ‘free speech’. The AI predicted an 80% probability of success at trial.⁵

The AI also analysed Judge McMahon’s historical scepticism toward DMCA claims [‘digital copyright infringement allegations’].

The AI model recommended emphasising, before the judge, statutory interpretation over technical arguments.⁶

In November 2024, that strategy secured the dismissal of major claims.

Three, AI is transforming judicial decision-making itself

Some courts in Shenzhen, China, have now established the world’s first court that systematically uses large language models (LLM). These AIs generate judicial opinions in most types of civil and commercial cases, including complicated contracts and corporate disputes.⁷

Based on these initial drafts, the judges then write their own judgments: why can’t our judges do it, so long as they do it honourably, and within accepted protocols?

[3]. Legal Writers and Researchers: AI’s Impact

For legal writers and researchers, AI presents both an unprecedented opportunity and significant risk. AI now generates first drafts in moments based on simple prompts. This allows writers to focus on higher-level analysis rather than routine drafting.

Legal research, once painstakingly slow, has been revolutionised. Attorneys using AI-powered research tools complete their work 24.5% faster than those using traditional methods. For the average attorney, this translates to saving between 132 to 210 hours a year.⁸. Imagine the old-style of writing, rewriting, and editing contract documents or arguments and summations for courts. That has changed.

[4]. Is AI merely making us faster?

Not at all. It is making us better. Not cleverer, but quicker. Nearly half (45%) of US attorneys believe they would have missed important precedents without the help of AI. By analysing vast datasets, AI makes connections that human researchers might overlook.⁹

And yet, our professional and ethical responsibilities remain our guiding light.

[5]. The Dangers: The consequences of uncritical AI use are severe

Hallucinations are fake citations or fake arguments.

Two bodies conducted a recent study on this: Stanford RegLab [research laboratory for effective governance] and HAI researchers [Human-centered AI experts.¹⁰ The research concluded that even specialised legal AI tools still produced incorrect information at alarming rates. Lexis+ AI and Ask Practical Law AI systems produced incorrect information more than 17% of the time.¹¹

Just a fortnight ago, attorneys using established legal research AI faced judicial wrath for ‘fake’ citations. In a recent Wyoming federal case against Walmart, lawyer Rudwin Ayala appeared after being granted leave for “temporary court admission”. It is called the “pro hac vice status”. He had drafted motions with AI-fabricated citations. The judge withdrew Ayala’s “pro hac vice status” [temporary court admission for a specific case] and the court fined him $3,000.¹²

In at least seven documented cases across the U.S., courts have questioned, reprimanded, or sanctioned lawyers for relying on AI-generated misinformation in legal filings. These infractions will only grow with time; unless ground rules are laid out.

[6]. What does all this mean for our future?

AI will not replace us. It will redefine us. It will not make you cleverer, just more efficient – perhaps even more precise. It will make your written work more readable, and certainly, shorter.

But institutional resistance is always there. Only demonstrable results will change minds. For example, the senior partners of a Bristol law firm initially resisted AI. After AI adoption and their firm’s productivity was measured, the results showed a 43% increase in associate productivity. The scores on mental well-being showed a 20% improvement.¹³

Some people fear this change.

Others embrace it blindly.

Both are wrong.

[7]. The Greatest Threat to Litigators

In my decades of practice, I have witnessed a troubling spectacle. Counsel copy vast citations, like scribes of old. They plaster these across their arguments, without discrimination. The right cases appear alongside the wrong ones. Obiter dicta masquerade as central holdings.

Some counsel ignore adverse authorities entirely. This stems not from dishonesty, but from pure ignorance: as Shakespeare wrote, “There is no darkness but ignorance.”

[8]. The Commonwealth Burden

When I began practice, a few journals sufficed. American attorneys rely solely on their own precedents. English barristers need only English law.

But we who practice in former colonies face a different fate.

Australian and New Zealand lawyers must navigate three jurisdictions.

Malaysian, Singaporean, and Indian advocates suffer the most. Because of their colonial past, counsel must trawl through Indian, British, Australian, and New Zealand authorities. What foreign case you cite depends on which foreign the legislation you rely on.

The web grows ever more complex.

I wonder: do our London colleagues understand the burden of lawyers at Kuala Pilah, or Alor Setar?

[9]. The AI Revolution

AI will level this uneven playing field. Complex research will become simple. Citations will flow like water.

But one caveat remains supreme: ethics.

[10]. The Duty of Candour

The greatest threat to litigators is not complexity, but the erosion of the very truthfulness upon which justice depends..

Lawyers serve as officers of the court. We bear a sacred duty to speak the truth. Hamlet’s words ring eternal: “This above all, to thine own self be true.”

Mistakes happen: wrong cases get cited, ratios are misunderstood: because these errors come with the territory, the courts understand human frailty.

But knowingly citing false authorities crosses a line. It opens counsel to misconduct charges, and worse still, contempt of court. When deception is discovered, courts act suo motu (of their own initiative and without complaint).

[11]. The Thin Line

A gulf separates honest error from deliberate falsehood. Contrast these situations: some counsel send copious authorities to court, and most go unused. This reflects not dishonesty, but poor judgment.

Counsel must verify every citation; and check every authority’s relevance.

AI amplifies this risk.

[12]. Tales of Mishap

How often does the following scene unfold? A junior prepares brilliant arguments. The night before the hearing, you discover a fatal flaw. An appellate court has overturned your foundational case, or a crucial principle has been modified.

This is neither dishonest nor negligent – for busy counsel rush against time, and often under terrible stress. Coronary disease is peculiar to litigation lawyers. It hounds them like a dark shadow. Yet it is the kind of heedlessness courts frown upon.

A good counsel rises and withdraws the point, while some stay silent.

When the judges notice, a painful, public reprimand follows. The experience is excruciating to endure :and agonising to observe.

[13]. The Price of Error

Sometimes you become collateral damage. I recall a partner who bore punishment for a junior’s affidavit. He was overseas when the events occurred. The newspapers had a field day.

Imagine such drama on the AI front. As the Bard reminds us: “The truth will out.” In our digital age, it will emerge faster than ever before. And so the AI must shield us.

[14]. The Dangers that Judges Face

Instead of poring through reams of useless verbiage, a court can use AI to do the heavy lifting. It can conduct advanced reasoning. This cuts down litigants waiting time by 50%. It increases judicial productivity by about 40%. There was a time when the courts would not accept printouts from an electronic journal. Now they do.

In China, courts already use AI. But courts must stop relying on it at some point: then, judges must step in. The haunting question is: when?

One must change with times, but how far should a judge step?

[15].The Human Touch

Consider the delicate work of judges. They must weigh cause papers on the fine scales of law. They must listen to witness testimony.

Often, lawyers do not understand the daily judicial grind. Judges must sift and weigh evidence. They must draw accurate conclusions from copious documents or the demeanour of prevaricating witnesses.

These tasks are so complex that they demand a human mind: a machine cannot do that. But AI can read notes of proceedings, caselaw bundles, or reams of documents. It can answer pointed questions.

[16]. The Paradox of the Mind

The mind is a complex network of neural webs and memory. It can be fair, or partial. Undesirable elements often corrupt human judgment. Emotional ties cloud it. Latent prejudices lurk within it, or hidden religious beliefs may cast shadows.

In the darkest hours of history, the nobility of the mind shines brightest. As Aristotle wisely observed: “The law is reason, free from passion.”

[17]. The Cold Algorithm

Unlike AI’s cold, calculating algorithms, trained processes of fairness animate the judicial mind. Equity flows through human consciousness. No machine learning can match this sacred quality.

Any juridical analysis must spring from human thought. All reasoning must bear human fingerprints. Every decision must carry human compassion and understanding.

[18]. The Price of Automation

Judges who rely solely on AI court disaster. The consequences can be terrible: disciplinary proceedings await; dismissal looms. Like shadows at dusk, the dangers multiply.

The day is not far when a judgment may be set aside on the ground ‘AI use’.

That should not drive one to timidity. A judge must step carefully through this digital minefield.

[19]. Ancient Wisdom for Modern Perils

Navigating technological dangers requires no new skill. We have practised this art for millennia.

Watch the motorcyclist delivering food parcels. Observe the jogger threading through traffic. They all know the dangers. They all try to avoid them.

Some will fall: that is for certain.

True it is that the scales of justice still require human hands to hold them steady.

But wisdom lies not in avoiding all risk, but in walking the narrow path between progress and peril.

[20]. But humanity always learns, always evolves

There will come a day when a client walks into a booth, pay RM100.00 to lodge a complaint. A machine will tell him when, where and how he can get his remedy. It may be a ‘no lawyer, consumer court’. Those days are fast approaching.

The judiciary and the legal profession cannot stop the march of progress. Their fear is: will progress trample over our collective sense of justice: like so many tanks flattening a garden of roses?

[21]. What do judges, lawyers, and writers yearn for?

To think clearly.

To write simply.

To speak powerfully.

To avoid accidental omissions – or commissions.

But that is to walk the narrow – and often harrowing – path of justice. There is no way out: only through.

[22]. Tomorrow’s courts and lawyers will combine the profession’s timeless virtues and ethics, with AI-enhanced capabilities

Consider a scenario where judges and lawyers are transparent about their use of artificial intelligence. Suppose they operate within a clearly established ethical framework. And suppose they can adhere to proven judicial protocols. Under these circumstances, why would not the entire justice system to migrate to the use of AI? Why would one not capitalise on this extraordinary advancement?

One must avoid the tyranny of algorithms. Yet, the transformation of our legal institutions needs to harmonise both the Rule of Law and the Rule of Technology.

As long as the judiciary and legal profession remain anchored to these fundamental values, society’s pursuit of justice will endure.

The choice is yours — be you a judge or a litigator.

What will you do with it?

∞§∞

Gratitude:

The author thanks UK Menon, Gana Naidu, KN Geetha, TP Vaani, JN Lheela, and Lydia Jaynthi.

Acknowledgements: the image is from Getty Images, Unsplash

@Copyright reserved.

All content on this site, including but not limited to text, compilation, graphics, documents, and layouts, is the intellectual property of GK Ganesan Kasinathan and is protected by local and international copyright laws. Any use shall be invalid unless written permission is obtained by writing to gk@gkganesan.com

The Promise and the Perils of Artificial Intelligence in Court Work

You May Also Like