Artificial intelligence (AI) has many beneficial applications that can improve quality of life, but accountability and transparency can be elusive when problems arise. This paper looks at the tools, systems and measures needed to ensure that when AI systems do more harm than good, their administrators are held accountable. Trade secrets law protects the source code, algorithms and data sets upon which AI systems are built, but they also prevent those harmed by these systems from getting answers. Requests for disclosure, data sharing and due process are being denied in the name of trade secrets protection. Complicating the matter is that trade secrets are treated differently in various jurisdictions.
