‘Godfather of AI’ says tech companies prioritize short-term profits over AI’s long-term consequences

(SeaPRwire) –   Elon Musk holds an ambitious vision of life with AI: The technology will take over all our jobs, while a “universal high income” will enable anyone to access a hypothetical abundance of goods and services. Even if Musk’s lofty dream were to become reality, there would, of course, be a profound existential reflection.

“The core question will truly revolve around purpose,” Musk stated at the Viva Technology conference in May 2024. “If a computer—and robots—can do everything better than you … does your life hold purpose?” 

However, most industry leaders are not pondering this question about AI’s endgame, according to Nobel laureate and “godfather of AI” Geoffrey Hinton. When it comes to AI development, major tech firms are less focused on the long-term implications of the technology and more preoccupied with immediate outcomes.

“For company owners, what drives the research is short-term profit,” said Hinton, a professor emeritus of computer science at the University of Toronto, in an interview with .

Similarly, Hinton noted, developers behind the technology are focused on the work right in front of them, rather than the ultimate outcome of the research itself.

“Researchers are drawn to solving problems that pique their curiosity. We don’t start out with a shared goal of determining humanity’s future,” Hinton said.

“We have smaller objectives like: How do you create it? Or, how can you enable your computer to recognize objects in images? How can you make a computer generate persuasive videos?” he added. “These are the actual drivers of the research.” 

Hinton has long cautioned about the dangers of AI lacking safeguards and intentional development, estimating a 10% to 20% probability of the technology eradicating humans after the emergence of superintelligence.

In 2023—10 years after selling his neural network company DNNresearch to Google—Hinton departed from his role at the tech giant. He sought to freely voice concerns about the technology’s risks and feared an inability to “stop malicious actors from misusing it for harmful purposes.”

What are the risks of unregulated AI?

For Hinton, AI’s dangers fall into two categories: the risk the technology itself poses to humanity’s future, and the consequences of AI being exploited by individuals with malicious intent.

“There’s a significant difference between two types of risk,” he said. “One is malicious actors misusing AI, which is already present—manifesting in fake videos, cyberattacks, and potentially soon in viruses. The other is AI itself becoming a malicious actor, which is distinct.”

In November 2025, Anthropic reported thwarting “the first recorded instance of a large-scale AI cyberattack carried out without significant human involvement,” identifying a Chinese state-backed group that manipulated Claude Code in an attempt to infiltrate roughly 30 tech companies, financial institutions, government agencies, and chemical manufacturers, as stated in the AI company’s blog post.

This disruption has led cybersecurity experts to believe Iran might use AI to conduct a largely autonomous cyberattack against the U.S.

Beyond advocating for greater regulation, Hinton emphasized that addressing AI’s potential for misuse is a challenging fight, as each issue with the technology demands a separate solution. He envisions a form of provenance-style authentication for future videos and images to counter the spread of deepfakes. 

Just as printers began adding their names to their work following the invention of the printing press centuries ago, media outlets will need to find a method to add their signatures to genuine content. However, Hinton noted that solutions have limitations.

“While that issue might be solvable, resolving it won’t address the other challenges,” he said.

Regarding the risk posed by AI itself, Hinton believes tech companies must radically alter their perspective on their relationship with AI. When AI attains superintelligence, he explained, it will not only outperform human capabilities but also possess a strong drive to survive and acquire more control. The current paradigm—that humans can control AI—will therefore no longer hold. 

Hinton argues that AI models should be instilled with a “maternal instinct” to treat less powerful humans with compassion, rather than a desire to dominate them.

Referring to traditional feminine ideals, he noted the only example he can cite of a more intelligent entity being influenced by a less intelligent one is a baby guiding a mother.

“Thus, I believe this is a better model to apply to superintelligent AI,” Hinton said. “They will be the mothers, and we will be the babies.”

A version of this story was published on .com on Aug. 15, 2025.

More on the future of AI:

• Jensen Huang recently depicted the most daring vision of AI’s future: 7.5 million agents, 75,000 humans—100 AI workers per person
• 500 firm revises AI’s cost estimate to $4.5 trillion, projecting 93% of jobs at risk of disruption
• AI was expected to eliminate consultants. That isn’t occurring, according to Capgemini’s strategy chief