Cheaper and quicker
The public is likely to become comfortable with machines deciding some disputes as the technology becomes more reliable, the Master of the Rolls has predicted.
Sir Geoffrey Vos said that litigants may eventually choose to have cases determined “in minutes by a machine at little or no cost” rather than wait months or years for a human judge to hand down a decision.
Speaking at the John Lehane Memorial Lecture at the Supreme Court of New South Wales, the top judge said he believes litigants will “come to have confidence in machine-made decisions” once they are shown to be as reliable, or more reliable, than decisions made by humans.
“Machine-made decisions will, of course, be cheaper and quicker to obtain,” Vos said.
He suggested the shift could begin in areas where outcomes can be determined largely through algorithms, pointing to the assessment of damages in personal injury and medical negligence claims, as well as share valuations in minority shareholder disputes.
Vos also noted that machine-led dispute resolution is already a reality in some settings, pointing to eBay, which he said decides 60 million disputes on its platform each year, with 90% resolved without human intervention.
But the Master of the Rolls drew a line between parties choosing an AI decision-maker and courts imposing one on them. “Consensual machine-made decisions are one thing, and non-consensual ones are quite another,” he said, arguing that there are “important reasons of legal principle” why courts may be unable to force machine-made judgments on unwilling parties.
Vos said the right to a human judicial decision is “deeply entrenched” and argued that judges should not be replaced without legislative change.
He also warned that the line between an AI making a decision and merely assisting a human judge could become increasingly difficult to maintain as the technology improves.
“If a judge is presented with 1,000 decisions, economic and time pressures will inevitably lead to judicial approval of the machine’s recommendations,” he said.
Despite those concerns, Vos predicted that acceptance of automated justice is likely to grow, particularly at the lower-value end of the court system. “It is quite likely, I think, that humans will come to accept, on economic grounds, machine-enabled resolution of small disputes at first instance,” he said.
The comments came as part of a wider speech in which Vos urged lawyers and judges to embrace AI, while maintaining a red line against fully autonomous judicial decision-making without the “freely and fully informed consent” of the parties.
Interesting timing been building something that solves exactly this for UK law firms. Happy to share if anyone wants to take a look.
We don’t give a sh*t. Thanks.
Lol
If the woeful state of most customer-assistance chatbots are anything to go off, decent AI legal decisions are a long, long way away.
He has been championing AI, like Susskind, for many years. I do wonder why there’s an intensity to put people out of jobs. They’ll dress it up as “efficiency” and “reducing backlogs”, but the reality we all know is that the MoJ has to cut its budget year on year, and AI makes big tech profitable. In the end its the working people who are losing out – both those making decisions and everyone who relies on working within the judicial system (admin, staff etc) and also the litigants. It’s a scary time when big tech and those championing it, who, lets face it, are already comfortable, keep pushing AI our way.
The simple fact of the matter is they don’t think it will affect them. The vision here is that the small “unimportant” cases will be dealt with by AI, but the significant ones will still go to well-respected judges with lengthy careers. Quite how any future judges will gain experience on the bench and demonstrate a record of sound decisions if the computers are doing it, I don’t know.
The AI bubble will burst and then we’ll see where we are. Far from AI making big tech profitable, AI is actually an alarming money-sink, with the biggest AI company in the world (OpenAI) losing tens of billions (yes, with a B) every year. We’re funding AI largely just because China is.
Judges don’t necessarily start at the bottom of the judicial ladder and work their way up. A lot of High Court Judges have no judicial experience before becoming a High Court Judge (or Deputy High Court Judge). Instead, they have a lot of experience advocating in the High Court.
Why is it always the older generation that become enamoured with superficial tech developments? They’re also the most prone to various online scams… Don’t get me wrong – AI is a big leap forward but it’s nowhere near what 60yo partners and judges who can’t convert pdfs claim it is.
Yet again senior judiciary contemplating what more can be hacked off of the legal system instead of lobbying for funding.
It’s all on the brink of collapse, and then where will we be?
I can see AI judges being very popular in small claims court or anywhere else where litigants are typically unrepresented. The AI will be able to spend much more time with the litigant. They can take an inquisitorial approach and interview each party to get all the information they need. That will be much less intimidating for litigants than being expected to present a case to a human judge when they have no idea what’s actually important. It will probably lead to better decisions, since poor presentation won’t get in the way of justice.
I think we’re a long way from society being ready to let AI decide important cases, no matter how capable it may be.
Have you ever dealt with LiPs in front of judges?
Hardly any LiPs can identify a relevant fact or know or care about the law. I foresee a lot of AI machines shouting at LiPs and then catching fire and billowing smoke as they realise the futility of giving them an important job managing the disputes of dunderheads.
That’s exactly my point. LiPs don’t know how to present a case to a judge. That’s why AI would be better. They wouldn’t need to present a case. The AI could interview them to determine the relevant facts. Judges don’t have time to spend several hours interviewing each litigant in a £1,000 case, but AI can take all the time it needs.
Okay, I, as a Claimant, uses A.I to “Help” formulate my Claim, it will clearly give me the ideal winnable scenario, now the Defendants do the same, uses the same A.I platform obviously it will advise the Defendants the best ways to defeat my Claim, despite the A.I platform originally told me I would win, ergo, which one will be right to make Judgement 🤔
Why would you assume the AI will tell you you will win? It will tell you how best to present your case. Presenting you case optimally doesn’t necessarily mean you will win. If you have a weak case then you have a weak case, regardless of how you present it.
You forget that AI has a tendency to hallucinate and completely make stuff up just to appease the questioner, so it is actually more likely that it will digitally nod and bow to whatever a claimant/defendant wants to argue (making up laws and case references along the way, as seen in real world cases where AI has been used to prepare court materials)…
In short, even if you do have a weak case, AIs can often decide your case is the best in the world, just to keep you chatting to it! It is algorithmically inclined to lean into the confirmation bias that us humans love.
Free chatbots like ChatGPT have that tendency, since it serves their purpose as a marketing tool. That’s not a tendency of AI generally. Nobody is suggesting that ChatGPT judge cases. It would be AI that is specifically designed for that task.
Maybe so, but in the spirit of “cheaper and quicker”, the government/MoJ cannot be trusted to design an AI of such a calibre. It will cheap-out/be under-resourced and thus unable to provide a quality AI, just look at the NHS.
When they are trying to get rid of juries, it seems worrying that they may one day get rid of judges too! Imagine walking into a court with nothing but robots determining your fate…
They’ll outsource the work to a tech company. They’ll be a very high profile client, so the tech company will have a strong incentive to do a good job.
I think I would have more confidence in even a cheap AI than a typical juror…
Alas I have less confidence in government procurement than in a typical juror, and I’d much rather we keep people employed in legal roles than fill them with bots!
It’ll be the age-old battle of a strong incentive to do a good job vs a strong incentive for private companies to profit as much as possible while sacrificing quality in the process…
Valid points from both of us, who knows how it will play out!
Vos’s intellectual stature matches the physical kind
Integrity has left the building
Why don’t judges – particularly the very senior ones – want to judge? It’s always about getting litigants out of court. Do your jobs.
Ai wasn’t meant to put dogs in Van Gough’s paintings!