Expert urges action on AI risks that could threaten humanity

Artificial intelligence could eventually pose a threat to humanity if it is not handled properly, cybersecurity expert Marcus Murray told SVT. He said warnings from AI leaders should be taken seriously and used to guide preventive action.

Geoffrey Irving, who has worked at companies including OpenAI, wrote in an open letter over the weekend that he believed there was about a 50 percent risk that people would die because of the development of AI systems smarter than humans. He said actions taken over the next two to 10 years would determine the outcome.

Murray said he could not assess whether people would die in the future, but argued that the risks needed to be discussed. Many important systems, including those used in healthcare and weapons, are connected to computers. AI has already hacked such systems, he said.

He described AI development as exponential. At Anthropic, AI accounts for 80 percent of the capacity to develop new code, according to Murray. He said AI is now building AI, raising questions about what these systems will be designed to do.

Murray said AI could cause serious harm to humanity in the future, but that this could be avoided if the technology were managed properly. He called for regulation to steer its development and ensure it proceeds safely.

Tags: artificial intelligence, cybersecurity, technology