Could artificial superintelligence spell the end of humanity?

Growing technology is causing growing concern

Robot with abstract thoughts
Artificial intelligence could pose a threat to humanity if it cannot be controlled
(Image credit: Vertigo3d / Getty Images)

Many experts expect artificial intelligence to evolve into artificial general intelligence (AGI) and artificial superintelligence (ASI) that would surpass humans in most capabilities, and it may be here sooner than we think. ASI is the "most precarious technological development since the nuclear bomb," said former Google CEO Eric Schmidt, Scale AI CEO Alexandr Wang and Center for AI Safety Director Dan Hendrycks in a newly published paper.

'Logic that we can't foresee'

Even if humans use means to limit ASI, a superintelligent system "could use trickery, lies and persuasion to get people to let it out," said The New Yorker. To prevent this, AI would need to be trained to prioritize human-like morals and ethics. This raises other concerns because "aligning an AI with one set of values could knock it out of alignment with everyone else's values," and lead to bias and discrimination. Also, "how do you tell the difference between a machine that's actually moral and one that's simply convincing its trainer that it is?"

The Week

Escape your echo chamber. Get the facts behind the news, plus analysis from multiple perspectives.

SUBSCRIBE & SAVE
https://cdn.mos.cms.futurecdn.net/flexiimages/jacafc5zvs1692883516.jpg

Sign up for The Week's Free Newsletters

From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.

From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.

Sign up

The threat to humanity is not just from AI but also from other humans. "A nation on the cusp of wielding superintelligent AI, an AI vastly smarter than humans in virtually every domain, would amount to a national security emergency for its rivals, who might turn to threatening sabotage rather than cede power," Schmidt and Hendrycks said at Time.

'A fundamental difference between human intelligence and current AI'

The advancement of AI is not a problem in and of itself. "The potential benefits of superintelligent AI are as breathtaking as they are profound," said Forbes. "From curing diseases and reversing aging to solving global warming and unlocking the mysteries of quantum physics, ASI could help us overcome humanity's greatest challenges." All that is needed is preparation. "This means investing in AI safety research, developing robust ethical frameworks and fostering international cooperation to ensure that superintelligent systems benefit all of humanity, not just a select few."

However, the threat of ASI may not be as imminent as some believe. "We expect AI risks to arise not from AI acting on its own but because of what people do with it," said Wired. AI researchers have "come around to the view that simply building bigger and more powerful chatbots won't lead to AGI." Also AI is "far from achieving the kind of versatility humans demonstrate when learning diverse skills, from language to physical tasks like playing golf or driving a car," said IBM. "This versatility represents a fundamental difference between human intelligence and current AI capabilities."

"The reality is that no AI could ever harm us unless we explicitly provide it the opportunity to do so," said Slate. "Yet we seem hellbent on putting unqualified AIs in powerful decision-making positions where they could do exactly that."

Devika Rao, The Week US

 Devika Rao has worked as a staff writer at The Week since 2022, covering science, the environment, climate and business. She previously worked as a policy associate for a nonprofit organization advocating for environmental action from a business perspective.