(NewsNation) — As artificial intelligence continues to advance at a rapid pace, one AI safety expert is warning the technology could become increasingly difficult to control without stronger safeguards.
Brendan Steinhauser, CEO of the Alliance for Secure AI, told “Morning in America” while he is excited about AI’s potential to benefit humanity, he is concerned the technology is advancing faster than researchers can solve critical safety and control issues.
“I’m very excited about the technology and what it can do for humanity,” Steinhauser said. “I think we’re currently not on the right path. We don’t have the proper safeguards in place.”
Meta’s new AI model can be downloaded on laptops, phones
Steinhauser said researchers have not yet fully solved what is known as the AI “alignment” problem — ensuring capable AI systems remain under human control and behave as intended.
He said the concern is becoming more urgent as AI models continue to improve.
According to Steinhauser, AI capabilities are increasing every four to five months, raising questions about how humans will maintain control as systems become substantially more powerful.
“It’s going to be a real problem as they get more and more capable, hundreds of times smarter and faster, and ultimately thousands of times faster and smarter than human beings, he warned. “And so I think we have a real problem on our hands.”
Bernie Sanders says AI companies need to pause to ‘avoid disaster’
Steinhauser argued developers should be cautious about building increasingly advanced systems approaching artificial general intelligence, or AGI, until researchers are confident they can safely control them.
The debate over AI safety comes as companies continue to develop more powerful models, with experts divided over how quickly the technology should advance and what safeguards should be required.