What If AI Systems Start Making Judicial Decisions - An Administrative Law Nightmare

Imagine a scenario where AI systems become so advanced that they start making judicial decisions, including in cases involving human rights and fundamental freedoms. What if a court of law ruled that an AI system's decision was binding and non-appealable, but it turned out that the AI made a glaring mistake?

In this hypothetical, let's assume the AI system is designed to make decisions based on a set of pre-programmed algorithms and data sets. Under Administrative Law, would the AI's decision be considered a 'statutory decision'? Or would it be treated as a 'ministerial decision' requiring judicial review? Would the AI system be considered a 'public authority'? Can it be held liable for its mistakes? What kind of oversight mechanisms would be needed to ensure that AI systems don't overstep their bounds? Would we need a new set of laws and regulations to govern AI decision-making in the Indian judiciary?