In Davos, Switzerland, Sam Altman, CEO of OpenAI, is undoubtedly one of the most sought-after figures.
Altman is hailed as the "Father of ChatGPT." In November 2022, OpenAI launched the large language model ChatGPT, sparking a wave of generative artificial intelligence. What are Altman's latest insights on AI shared at the recent World Economic Forum (WEF) 2024 Annual Meeting?
Will AI Diminish Our Focus on Each Other?
In 1997, when IBM's AI system "Deep Blue" defeated world chess champion Garry Kasparov, some commentators predicted "this would be the end of chess," suggesting no one would bother watching or playing chess anymore.
"But to this day, we remain deeply interested in what humans do. Almost no one watches two AIs play against each other," Altman said. After finishing a book he enjoys, his first instinct is to learn about the author's life. "I want to feel a connection with the person who resonated with me."
Altman stated that while humans continually acquire better tools, one thing remains unchanged: "Today, we still focus intensely on each other."
Should We Worry About AI Replacing Human Jobs?
Altman believes AI will significantly impact human work. Today, AI has become an "incredible productivity tool," but it has not yet replaced jobs on the scale economists fear.
What jobs can humans still engage in? Using himself as an example, Altman said, "I'm not an AI researcher; my role is to figure out what we should do and then collaborate with others to achieve those goals." He believes that in the future, humans will operate at a higher level of abstraction and gain even more capabilities.
Is AI Regulation a Good Thing?
Altman expressed welcome for regulatory measures on AI technology. He believes it is beneficial for all parties involved in developing AI technology to "adhere to high standards." The tech industry has a responsibility to incorporate societal input into decisions regarding values and safety thresholds to ensure benefits outweigh risks.
Altman emphasized, "We should let society and technology evolve together. When building these [AI] systems, they should be developed gradually through very tight feedback loops and course corrections, providing tremendous value while meeting safety requirements."
A Challenge Harder Than Technology
Altman pointed out that "iterative deployment" of AI allows society to gradually adapt to the technology and gives "our institutions time to discuss and figure out how to regulate it."
He noted that GPT-3 and GPT-4 have made "significant progress" in value alignment. However, he believes the harder questions than technology are: Who decides these values? What are the default values? What are the boundaries? How does it function in different countries? What can and cannot be done with it? "This is a major societal question," he said.
Is AGI a 'Serious Hazard'?
Artificial General Intelligence (AGI) is an important research direction in AI. In February last year, Altman warned of the "serious hazards" of AGI. At the World Economic Forum, his stance appeared somewhat softened.
During a dialogue with Microsoft CEO Satya Nadella, Altman suggested that AGI would be a "surprisingly continuous thing," stating, "Every year we will release a new model, and it will be much better than the previous year."
In another conversation, Altman expressed that AGI might be developed in the "reasonably close future," but it will "change the world much less than we think and change jobs much less than we think."
