The German Jan Leike is trying to prevent a super AI from turning against humanity at the chat GPT builder Open AI. To be on the safe side.
Talking to Jan Leike is like calling on a spaceship. It’s not easy to penetrate the mind of someone who spends their day thinking about the question: How do I stop artificial intelligence from going crazy? And who thinks about each of his sentences for a long time before he says it. They’re on guard for Open AI. The company Leike works for has aroused public interest and skepticism like no other in recent years. Open AI has started a hype about generative AI with Chat-GPT. The chatbot generates human-looking texts on command. Leike, a German, is an expert in machine learning. At Open AI he is one of the leaders teams for the so-called superalignment (Alignment stands for the alignment of a possible superintelligence with the interests of people). Leike’s job description sounds like science fiction to many: He is supposed to ensure that computer superintelligence – which critics consider impossible – one day serves people instead of the other way around. At Open AI they claim that superintelligence will come at some point; after all, they are working on it themselves.