Microsoft AI chief says AI is not conscious and we should keep it that way

Some AI researchers are questioning whether AI could one day acquire consciousness. Not if Microsoft's AI CEO has item to say concerning it.

Mustafa Suleyman stated in a lengthy X article on Sunday that AI models currently don't have awareness and that developers should keep it that way.

"If it is possible, according to everything I know, evolving conscious AIs would be reckless, and we should not build them," Suleyman said. "If awareness started to appear in training, I would call for everyone to halt AI betterment until we could scheme out those properties."

If any evidence were to appear that AI models were conscious, Suleyman stated it must be disclosed to the broad community immediately so we can decide our next steps as a society.

"We AI developers have no correct whatsoever to bring a new category of conscious beings into the world, equal if several group may desire to do it alongside AI benefit as a top priority," he said. "The another 8 milliard conscious individual beings on the earth have not signed up to this test and their ethical rights, safety, and benefit are what we have an duty to defend archetypal and foremost."

Conversations about whether AI models could one day rotate into conscious are picking up steam in the tech world. Last April, Anthropic started a investigation program to examine and prepared to navigate "model welfare." In a blog post, the startup questioned whether developers have to be concerned concerning models' possible awareness and experiences.

Google DeepMind published investigation document this twelvemonth concerning AI awareness and how ascribing that trait could current challenges.

On X, Suleyman stated that if an AI example acts as if it suffers, or genuinely believes it can suffer, it volition apt prioritize its own needs complete operational objectives. As a result, they could damage humans since they're competing for resources, specified as compute and energy, to prosecute their own goals.

"We should lone scheme AIs to so they are not conscious and not deserving of rights, and alongside the sole intent to accelerate individual wellbeing without being driven by their own intrinsic motivations.. We could afterward see whether this is easier or harder to align and keep safe," he said.