I Have a Dispute to Mediate -- Should I Use Artificial Intelligence as My Mediator?

We are all hearing about artificial intelligence—the exciting ways it can process a lot of data quickly and the troubling ways it can use bad data just as easily as good data, as well as create “hallucinations” based on misperceiving what are facts and what are not.

Still, with its reported efficiency and accessibility, one might wonder whether trying AI as a neutral to resolve a dispute might be better than using a human mediator.  AI does not appear to possess some skills, however, that seem essential to serving successfully as a mediator -- at least not yet.   

Lack of Empathy - AI cannot understand and empathize with emotions. Often in a dispute a party first needs to feel heard and understood about why the dispute exists before the party is ready to discuss solutions. Part of being able to consider the other side’s perspective and considering compromises, is first having confidence that one’s own perspective is heard and appreciated. 

Thinking Inside the Box - AI works on predefined algorithms. Although we can input facts to get AI to develop options, AI may not be able to think outside of those defined limits, to ask further questions to understand whether a fact not yet shared might expand potential solutions to a dispute.

Limited Creativity - AI has also been shown to be less creative than humans. Hence, while all the information and issues presented in a dispute may have been presented in monetary terms, for example, only a human mediator might consider a completely different means of settling the dispute -- such as an apology and some collaborative endeavor that hasn’t been explored yet and benefits both parties. This solution might be something that was only raised by one party’s off-hand comment during their discussions. AI might have dismissed the comment as irrelevant.

Can’t Build Trust and Create Calm - another aspect to a successful mediation is often that the parties trust the process and the mediator to protect the process. Often a party is comforted by the mediator’s commitment to stay with the parties until they have explored all viable solutions to their dispute, for example. AI is unlikely to make such a commitment.  Sometimes, it is simply the calm and confident demeanor of the mediator that allows a person to stay calm and focused and consider the other party’s ideas, thereby leading to a settlement. At this point, I doubt an AI mediator would generate that kind of calm and focus in a party.

What Happens to Confidentiality and Privilege? - There is also another concern with using AI as a mediator: the potential for risk to the confidentiality and privileged nature of the process that is otherwise available to mediating parties in the U.S. and other countries. Disclosure of information to AI may lead to waiver of those rights because AI generally uses the information received as data for improving its process. It integrates the information into its database. This “disclosure”, if authorized by the parties, may be viewed as a disclosure to a third party waiving all rights to protect such information and to keep it from discovery in a court proceeding in the unfortunate event that the mediation fails. Courts will no doubt be evaluating this issue around the world; it’s a concern for now.

So should we completely forget about AI? No. It can be useful in a mediation. A mediator and the parties can use AI to develop options, sort data, and organize issues, for example. It can also help locate outside information the mediator seeks (since it is built into most search engines nowadays). Perhaps one day AI will be a better choice, but for now, a human mediator, equipped with AI, seems the better choice to resolve our human differences.