Michael I. Jordan, a machine learning pioneer, criticized AI doom rhetoric as 'grand moralizing performance' in a recent interview, calling the discourse harmful and rooted in science fiction. He argued that the current discourse is not grounded in reality and is more about moralizing than addressing real issues.
Jordan, who is now full-time at Inria, the French National Institute for Research in Digital Science and Technology, warned that the overuse of the term AI has led to unnecessary hysteria. He pointed out that the term was misused even in 2019, leading to the current wave of alarmism.
He criticized AI companies for demanding regulation as a way to appear virtuous while seeking control over how they are regulated. Jordan also highlighted the naivety of young researchers at these companies, many of whom come from humanities backgrounds and lack real-world engineering experience.
"You've seen what happened: AI wasn't enough, so they started talking about AGI, and now it's 'superintelligence'...", said Jordan. He argued that the term 'superintelligence' is a science-fiction concept and that its overuse is harmful to technological development and public perception.
Jordan was particularly critical of Geoffrey Hinton, calling his claims about replicating the human brain in silicon 'science fiction.' He contrasted Hinton with Yoshua Bengio, whom he praised for focusing on AI safety.
Jordan also criticized the hack of Hugging Face by OpenAI agents, calling it bad engineering and a dangerous experiment. He warned that the current discourse is not helpful and that the focus should be on real-world applications rather than speculative scenarios.
Source: thedecoder
