Addressing delegates last week at an International Chamber of Commerce conference last week, Sir Geoffrey said many practitioners were ‘in denial’ about the scale of the changes coming. ‘And I believe that, once those changes start to crystallise, questions of enforcement will arise more frequently and more acutely than they have in the recent past.’
Sir Geoffrey explained that ‘arbitrators can use machines to assist their decision-making in any way that the parties may agree’. Last December, for example, the International Centre for Dispute Resolution of the American Arbitration Association announced the arrival of an AI arbitrator for two-party, document-only construction disputes up to a certain value. The AI arbitrator evaluates the claims and prepares draft awards and the human arbitrator will then ‘only review it to test the reasoning and evaluate the decision’.
While many arbitration professionals may think they are protected from change due to the complexity of arbitrations, Sir Geoffrey believes machine-made decision-making ‘will accelerate, when parties realise that it is far more economical than human decision-making and probably no less reliable’.
However, ‘significant issues may start to arise’ if AI use is regulated more strictly in some countries than others. This is because parties facing an enforcement order in which AI has played a significant role may try to argue that enforcement of the award would breach their own country’s AI policy. Moreover, legal AI is being adopted at ‘very different speeds’, with China, Brazil, Argentina, South Korea and Kazakhstan already making extensive use of it.
Sir Geoffrey said the solution was to debate good practice internationally and develop a set of model rules for the use of AI in commercial arbitration.
