Former Google CEO Warns: Autonomous AI Could Present an Existential Threat—and It’s Imminent

Eric Schmidt, the former CEO of Google, remarked that the emergence of autonomous artificial intelligence (AI) is imminent, and it could represent a significant existential risk for humanity.

“We are on the brink of having computers operate independently, making their own decisions,” Schmidt stated during his appearance on ABC’s “This Week” on December 15, where he has frequently highlighted the risks and advantages that AI entails for society.

“This is a perilous stage: When a system can improve itself, we must seriously contemplate disconnecting it,” Schmidt cautioned.

Schmidt is not the first technological leader to voice such worries.

The rapid ascent of consumer AI technologies, such as ChatGPT, has been extraordinary over the last two years, showcasing significant advancements in language processing models. Other AI platforms have also become increasingly proficient at producing visual artwork, photographs, and extended videos that, in many instances, closely resemble reality.

For some, this technological advancement evokes images from the “Terminator” series, which portrays a dystopian future dominated by AI and resulting in catastrophic outcomes.

Related Stories

Despite the anxieties surrounding ChatGPT and comparable tools, the consumer AI technologies currently available still fit into a category experts label as “dumb AI.” These systems are based on extensive datasets but do not possess consciousness, sentience, or the capability to act independently.

Both Schmidt and other specialists are not particularly concerned about these simpler systems.

Their focus lies on the more sophisticated AI—referred to in the tech community as “artificial general intelligence” (AGI)—which encompasses much more intricate systems potentially possessing sentience, along with the ability to develop independent motives that could be harmful to human interests.

Schmidt mentioned that such advanced systems do not presently exist, and we are swiftly approaching a new category of AI: one that lacks the sentience classifying AGI yet can still function autonomously in areas like research and military applications.

“In my 50 years in this field, I’ve never witnessed innovation on this scale,” Schmidt remarked about the rapid advancements in AI complexity.

He noted that further developed AI could yield numerous benefits for humanity, while also posing significant risks, such as the development of weapons and cyberattacks.

The Challenge

The challenge, according to Schmidt, is complex.

At a fundamental level, he reiterated a common viewpoint shared among tech leaders: should autonomous AGI-like systems become inevitable, it will necessitate extensive collaboration between corporate entities and governments worldwide to avert potentially disastrous outcomes.

However, achieving this is easier said than done. AI could potentially provide U.S. rivals, including China, Russia, and Iran, a competitive edge they might not otherwise gain.

Additionally, within the tech industry, there is currently intense competition among major companies—such as Google and Microsoft—to surpass each other, which in turn raises inherent risks regarding the security measures needed to handle a rogue AI, Schmidt pointed out.

“The competition is so fierce that there’s a concern one of the companies might skip essential safety steps and inadvertently unleash something harmful,” Schmidt stated, noting that the consequences would likely only reveal themselves post-factum.

On an international level, the challenge intensifies, with adversarial nations likely to perceive this new technology as revolutionary in their efforts to challenge U.S. global dominance and expand their influence.

“The Chinese are astute and recognize the value of a new kind of intelligence for enhancing their industrial strength, military power, and surveillance capabilities,” Schmidt explained.

This situation creates a dilemma for U.S. leaders in the tech sector, who must navigate existential anxieties about humanity alongside the risk of the United States falling behind its adversaries—a scenario that could adversely affect global stability.

In a worst-case scenario, such systems might be exploited to create devastating biological and nuclear weapons, particularly by extremist groups like ISIS.

Thus, Schmidt emphasized the need for the United States to continue advancing in this area to maintain technological superiority over China and other hostile states and factions.

Industry Leaders Demand Regulation

Schmidt criticized the current regulatory landscape as inadequate, expressing his expectation that government focus on enhancing technological safeguards would significantly increase in the coming years.

When anchor George Stephanopoulos asked whether governments were doing enough to oversee AI development, Schmidt replied, “Not yet, but they will, because there’s no other option.”

Though there has been some initial engagement in the arena—including hearings, legislative proposals, and other initiatives—during the current 118th Congress, it appears that this session will conclude without any significant legislation concerning AI.

President-elect Donald Trump has also underscored the considerable risks associated with AI, stating during an appearance on Logan Paul’s “Impaulsive” podcast that it is “extremely powerful stuff.”

He conveyed the critical need to sustain competitiveness against adversaries.

“It presents challenges, but we must lead the way,” Trump asserted. “It’s going to happen, and in that case, we must stay ahead of China. China represents the primary threat.”

Schmidt’s views on both the advantages and challenges posed by the technology resonate with reactions from other industry leaders.