They compared nuclear power as another example of a “potentially existentially endangered” technology. There is a growing need for an agency similar in nature to the International Atomic Energy Agency (IAEA), the world’s nuclear watchdog.
Over the next decade, “AI systems are likely to exceed the skill level of experts in most fields and perform productive activities on a par with one of today’s largest companies,” the OpenAI team wrote. ing. “In terms of both potential good and bad, superintelligence will be more powerful than any other technology humanity has had to contend with in the past. But we need to manage risk to get there.”
The statement echoed Altman’s comments to Congress last week, in which the US-based company’s CEO also testified to the need for a separate regulatory body.
Critics have warned against trusting calls for regulation by tech industry leaders who want to profit from unbridled development. Some say OpenAI’s business decisions stand in contrast to these safety warnings. That’s because OpenAI’s rapid deployment has sparked an AI arms race, putting pressure on companies like Google parent Alphabet to release products while policymakers are still grappling with the risks.
The Washington Post hopes its supporters and critics will influence debates about technology policy, so few lawmakers have a deep understanding of emerging technologies and AI, and AI companies are widespread. I have previously reported that they are lobbying.
Others warn of the risk of hindering the United States’ ability to compete on technology with its rivals, particularly China.
In a memo, OpenAI leaders opposed the development pause, adding that “stopping the creation of superintelligence would intuitively be dangerous and difficult.” The benefits are so great that the cost of building them is decreasing year by year and the number of entities building them is growing rapidly. ”
In his first congressional testimony last week, Mr. Altman warned about how AI could “cause grave harm to the world,” and vowed the company would continue to roll out the technology.
Altman’s message of willingness to work with lawmakers was received relatively warmly in Congress, as countries, including the United States, recognize the need to work with them. We must work to support innovation while addressing technology that raises concerns about privacy, safety, headcount reduction, and misinformation.
New York University professor emeritus Gary Marcus, a witness at the hearing, stressed the “mind-boggling” amount at stake and said OpenAI is “indebted” to its investor, Microsoft. . He criticized the company for deviating from its mission to advance AI to “benefit humanity as a whole” without being constrained by financial pressures.
The proliferation of ChatGPT and generative AI tools that create text, images, and sounds have surprised users and heightened the urgency of the regulatory debate.
At the G7 summit on Saturday, leaders of the world’s biggest economies made international standards of AI progress a priority, but have yet to draw substantive conclusions on how to deal with risks.
The Biden administration has said it is making AI a key priority, but the U.S. has so far been more sluggish than other countries, especially Europe. Washington policymakers have failed to pass comprehensive technology legislation for years, raising questions about how quickly and effectively they can develop regulations for the AI industry.
For the time being, ChatGPT’s creators called for “some degree of coordination” among companies working on AI research in order to allow “superintelligence development” to be safe and “smooth integration of these systems with society.” rice field. For example, the companies may “collectively agree that the growth rate of AI capabilities on the front lines will be capped at a fixed annual rate.”
“We believe that people around the world should democratically decide the limits and defaults of AI systems,” they added, but “we still don’t know how to design such a mechanism.” I admit that.
Cat Zakrzewski, Cristiano Lima, and Will Oremus contributed to this report.
