Artificial intelligence will be able to create its own next generations by the end of 2028
This was reported by The Guardian, citing a speech given by Jack Clark, co-founder of Anthropic, at the University of Oxford.
He predicts that within 18 months, companies fully controlled by AI will emerge, generating millions of dollars in revenue, and by the end of 2028, AI systems will be able to develop their own next generations.
Clark noted that the technology still carries the potential for global existential risks. He said that commercial and geopolitical rivalry between different countries and market players is hindering the pace of development.
Meanwhile, Edward Harkort, director of the Institute for Ethics in AI, stated that algorithms performing an ever-increasing number of tasks on behalf of humans could lead to a decline in people’s ability to make decisions independently, and advocated for the development of models that require greater cognitive engagement from humans.
Clark’s most conservative scenario envisages the separation of the machine economy from the human economy, the automation of scientific research, and the creation of technology never seen before.
He urged society to prepare for future changes, as a lack of regulation on the spread of AI will force humanity merely to react to the consequences.
Follow us on Telegram