Directors’ defence of reliance on recommendations made by artificial intelligence systems

Artificial Intelligence systems are rapidly improving their capabilities, leading some companies to incorporate them into their business decision-making processes (even at the boardroom level). This paper analyses and compares three different approaches in respect of the directors’ reliance defence in corporate law and offers some tentative suggestions as to the applicable legal principles when directors rely on information and advice from an artificial intelligence system. It argues that a statutory reliance defence is not necessary, because whether or not it exists, reliance is always a relevant factor in the analysis of whether a director’s conduct meets the applicable reasonableness standard. Where the advice and information is offered by AI, certain issues with AI (such as non-transparency, explainability, and narrowness) may arise. This paper discusses how each of these problems may affect the reasonableness analysis and recommends boards consider using Algorithmic Impact Assessments to assess the suitability of implementing an AI system prior to relying upon its recommendations.