64 terabytes of pure empathy

As we all know, the science described as Artificial Intelligence (AI) concerns the endowing of computers with the capabilities of human intelligence. When I hear the term AI, I think robots. I also think about blockbuster movies that have robots taking over the world causing mass destruction. Scary stuff. But AI has just got a whole lot scarier! The website legalfutures.co.uk recently carried the headline ‘Here come the robot mediators’ and published an article entitled ‘Robot mediator settles first ever court case’. My very existence is under threat! The late Stephen Hawking said in an interview in 2016 ‘I fear that AI may replace humans altogether’. He may not have intended to include mediators, but things don’t look good!

The History Of Technology In Mediation Services
The use of artificial negotiation systems has gained some traction in recent years. Smartsettle is an example, using game theory to deliver solutions. Smartsettle was the system used recently to achieve what is believed to be the first settlement of an online dispute in England and Wales facilitated by computer algorithm. The dispute was over unpaid fees for a counselling course. Each side made blind offers by placing a flag along a slider multiple times until a deal was reached. Computer algorithms learn the bidding strategies of each side so they can be helped to reach a settlement. Fine! But what have robot mediators got that the human variety do not have?

Ok, maybe negotiation support systems can improve the speed and efficiency of dispute resolution in simple debt cases. Maybe a human element isn’t needed, although if I didn’t think money was owed but nevertheless wanted to get rid of a silly claim, I would probably want a little rant about it first. So, could I rant (sorry, vent) to a robot? Would I be told to ‘grow some bolts’, or ‘talk to the motherboard because my GUI (graphical user interface) ain’t listening’? Or would it be a robotic death-stare followed by a simple ‘guilty’. The human element of mediation, the intuitive skills that all good mediators possess, just can’t be taught, well certainly not to a computer. Or so I thought! In 2016, Hanson Robotics unveiled an award winning human-like robot that can apparently distinguish between different people, understand emotions from tone of speech and facial expressions, and react appropriately through its own speech, human-like facial expressions and body language. They now have a number of such robots. So, just maybe, there could be some empathetic interaction between a robot mediator and parties to a mediation. And maybe I could vent (although could I do so without the risk of being exterminated?).

There is no doubt that robots are becoming more and more human-like. Apparently, in years to come, they will be commonplace in our homes. Futurologists tell us they will be cleaning for us, cooking for us and may even have rights. One senior judge has even suggested that computers could, in the future, have distinct legal personalities and own property. (Hey, Mal-Colm Z type, would you mind terribly watering our plants whilst we are away?).

The Future Of Mediation
So robots can do all sorts of things. But mediating! Really? When Elon Musk predicted that ‘there will be fewer and fewer jobs that a robot can’t do better’, did he really intend to suggest that a robot mediator could out-mediate the human variety? No, I bet he didn’t, because computers go wrong all the time. Think of the damage that could be caused with a malfunctioning robot mediator, even one listed in the top tier of Legal 500 or Chambers. Party A: ‘we would like you to take this offer to the other side’. Robot: ‘computer says No’. What if a malfunction resulted in a far more serious outcome. Party B: ‘I think Party A are behaving outrageously and are not here in good faith’. Robot to Party A: ‘you are terminated’. How many times have our computers suddenly started to work at half speed, or freeze? Would we really want our robot mediators saying ‘I will be back as soon as I have finished rebooting’ or ‘quiet please, I am analysing’. No, I don’t think we would.

Impartiality In Mediators
Malfunctioning aside, using any kind of algorithm in the context of dispute resolution introduces new issues. Impartiality or the absence of bias is important for a mediator. What can be less impartial or free from bias than an algorithm? Well, algorithms are designed by humans. Robots, or the algorithms that run them, will carry their own bias depending on who designs them. Indeed, there are some interesting issues emerging from predictive policing algorithms.

And will lawyers (until they too are replaced along with mediators, chefs, interpreters, surgeons etc) not seek to gain advantage by learning the various AI dispute resolution systems on the market, their idiosyncrasies etc and choose the most advantageous for their client? Would a lawyer who, for instance, has been on a training course run by a particular AI dispute resolution provider, gain an advantage when using that system over a lawyer who hasn’t taken the course?

And what if lawyers can’t agree on an AI system to use? Maybe they will need a (human) mediator to help them reach agreement!

AI is getting smarter and smarter as time goes on. The real danger though is when AI becomes better at (self) improvement or re-calibration than its creators and transcends the original purpose of its creation. The dangers of recursive self-improvement beyond the creator’s intentions maybe some years away, but what if robots could stop disputes arising in the first place? A truly horrific thought for everyone in the dispute resolution sector!

Trusted Experts In Mediation
Mediating is a tough job requiring a great deal of skill. Mediators must know when to listen and when to move things on. We all have our own individual style, but we must also be capable of style-drift depending on changing circumstances during the day. We counsel parties on the offers they make and responses to offers received, synthesising negotiating styles where necessary. We generate options for settlement, unlock fixed positions and overcome deadlock. We help parties think the unthinkable, help make solutions manageable. We don’t say things like ‘I am unable to comply’, ‘you don’t deserve to live’ or ‘guilty’, however bad a day we may be having.