Automated Systems and Artificial Intelligence
Mart Susi · Cambridge University Press eBooks · 2024
There are two human rights properties overshadowing academic discourse and policy documents in relation to automated systems and artificial intelligence which are operating in the digital realm. These properties are about the trust in artificial intelligence and automated systems, and secondly, of robustness. The latter appears in two aspects related to law governing the AI solutions and the AI itself. The degree of our trust in physical judges is related to our expectation that they are capable of distrusting the law. And our trust in automated systems and artificial intelligence is weaker because we assume that they completely trust the law. It could be said that we cannot find a better positivist than artificial intelligence. The moment artificial intelligence can independently exercise disavowal and intolerability formulas, that is, articulate a reasoned understanding that a specific legal norm or rule is profoundly unfair, then the non-coherence disappears. However, then we would no longer be dealing with artificial intelligence; perhaps we could term such a decision-maker an artificial brain.