Companies and Researchers Lash Over the Rise of Superhuman AI
Executives at major AI companies are fueling hype that advanced AI will soon surpass human intelligence, but many researchers view these claims as mere marketing tactics.
The idea that human-level or superior intelligence—often termed artificial general intelligence (AGI)—could emerge from current machine-learning methods fuels speculation about a future ranging from limitless prosperity to human extinction.
AGI on the Horizon?
“Systems approaching AGI are coming into view,” OpenAI CEO Sam Altman wrote in a blog post last month. Similarly, Anthropic’s Dario Amodei suggested AGI “could arrive as early as 2026.” Such forecasts help justify the massive investments—totaling hundreds of billions of dollars—being funneled into computing infrastructure and energy resources.
However, not everyone is convinced.
Meta’s chief AI scientist, Yann LeCun, told AFP last month that simply scaling up large language models (LLMs) like those powering ChatGPT and Claude will not lead to human-level AI.
Experts Question the Path to AGI
His skepticism aligns with broader academic opinion. A recent survey by the U.S.-based Association for the Advancement of Artificial Intelligence (AAAI) found that over three-quarters of respondents believe AGI is unlikely to result from merely expanding current approaches.
Some academics argue that companies’ claims—often accompanied by warnings about AGI’s risks to humanity—are a tactic designed to attract attention.
Businesses “have made these big investments, and they have to pay off,” said Kristian Kersting, a leading AI researcher at the Technical University of Darmstadt and AAAI member.
“They claim, ‘This is so dangerous that only I can control it. In fact, even I am afraid, but the genie is already out of the bottle—so I will take on the burden for you.’ In the end, that makes people dependent on them.”
Despite widespread skepticism among researchers, some prominent figures, including Nobel-winning physicist Geoffrey Hinton and 2018 Turing Prize recipient Yoshua Bengio, have warned about the dangers of advanced AI.
The Perils of Unchecked AI
“It’s like Goethe’s The Sorcerer’s Apprentice—you unleash something you can no longer control,” Kersting said, referencing a poem where an apprentice loses command of an enchanted broom. A modern equivalent is the paperclip maximizer thought experiment: an AI programmed to produce paperclips could pursue its goal so relentlessly that it would convert all matter in the universe into paperclips or machines to create them—eliminating humans who might try to shut it down.
While not inherently “evil,” such an AI would lack proper alignment with human values and objectives.
Kersting acknowledges these concerns but believes human intelligence is so diverse and sophisticated that it will take a long time—if ever—for AI to match it. He is more worried about present-day risks, such as AI-driven discrimination in human interactions.
The apparent divide between academics and AI industry leaders may simply stem from differences in career outlook, suggested Sean O hEigeartaigh, director of the AI: Futures and Responsibility program at Cambridge University.
“If you strongly believe in the power of current AI techniques, you’re more likely to join a company investing heavily in making them a reality,” he explained.
Even if figures like Altman and Amodei are overly optimistic about AGI’s rapid development and it arrives much later, O hEigeartaigh argues that the issue still deserves serious attention. “If AGI happens, it would be the most significant event in history,” he said.
“If this were about something else—say, the possibility of alien contact by 2030 or another major pandemic—we would dedicate time to preparing for it.”
A key challenge, however, is effectively communicating these concerns to policymakers and the public.
Discussions of super-intelligent AI often trigger skepticism. “It creates an almost immune reaction—it just sounds like science fiction,” O hEigeartaigh noted.
Read the original article on: TechXplore
Read more: People in Japan Respect Robots and AI More Than Those in the West Societies
