A former OpenAI expert who predicted that artificial intelligence would surpass humanity in 2027 has postponed his predictions to the 2030s, saying that technology is mired in the complexity of the real world.
Elchi.az reports that Daniel Kokotajlo, a former OpenAI employee known for his frightening scenarios about artificial intelligence bringing about the end of humanity, has updated his pessimistic predictions about the future of the world. Kokotajlo, who once caused a great stir in the technology world and pointed to 2027 as a turning point, admitted that the pace of artificial intelligence development is progressing “a little slower” than his initial predictions and has pushed back the crucial timeline for humanity.
Kokotajlo’s scenario, published in April last year and entitled “Artificial Intelligence 2027”, predicted the emergence of a superintelligence that would overthrow world leaders in the wake of uncontrolled development and ultimately destroy humanity. However, the complexity of the real world and the failure of artificial intelligence’s autonomous coding capabilities to fully accelerate progress led to a review of this dark prophecy. According to new forecasts, the self-improvement and transformation of artificial intelligence into superintelligence is now estimated for the early 2030s, not 2027.
Experts note that artificial intelligence’s success in narrow fields such as playing chess does not progress at the same pace in the face of the great inertia and complexity of the real world. The concept of “artificial general intelligence”, which was once close to science, is now slowly turning into a science fiction story for many researchers. Even OpenAI CEO Sam Altman’s cautious statements about his company’s goals, such as “we may fail completely”, prove that this great transformation in the technology world will not be as smooth as expected. As a result, the frightening possibility of artificial intelligence destroying humanity and turning the world into a solar panel farm is, for the time being, hampered by the bureaucracy and technical obstacles of the modern world.
Şayəstə