With concerns that China will overtake the United States in the Artificial Intelligence field, some experts are cautioning against enacting a pause, endorsed by tech leaders, on “out of control” AI systems.
In an open letter, hundreds of tech leaders, professors and researchers, including Elon Musk, advocated for an “immediate pause for at least six months” on all development of AI systems. They warn that the rapid advancement of contemporary AI systems poses a string of safety concerns, including the difficulty creators may have when it comes to understanding and predicting these new “digital minds.” But the letter received mixed reactions and some argued that pausing American innovation on AI gives China a competitive advantage in the field.
AI Race: China versus U.S.
“China is America’s most formidable competitors in the AI field,” Timothy Heath, a senior international defense researcher at Rand Corp, told Newsweek. “What makes them such a challenge for the United States is that they have enormous resource to put against AI and they have a different value set. This means China may be wiling to use AI in ways that threaten U.S. interests.”
While China is a global leader in AI research, its response to OpenAI’s new GPT-4 (Generative Pre-Trained Transformer), an even more powerful version of ChatGPT driven by a large language model (LLM), was seen as disappointing by investors. Baidu’s Ernie Bot, which was unveiled earlier this month, sent the Chinese company’s shares tumbling.
“Right now, major tech companies in China are behind U.S. tech companies on building and fielding large language models, but they’re certainly working hard on it as well,” Paul Scharre, the vice president and director of studies at the Center for New American Security, told Newsweek.
Open AI’s programs have both impressed and caused alarm with their ability to engage in human-like conversations and to excel in some of the most challenging standardized tests. The latest iteration can pass most Advanced Placement exams and score in the 90th percentile of the standardized bar exam for aspiring lawyers. As of Friday, Open AI’s CEO Sam Altman had not signed the open letter.
The concerns raised by programs like GPT-4 have prompted calls like the one in the open letter. Although many in the field recognize that halting research and development in favor of safety protocols in the U.S. is important, “China is not going to slow down its AI development in either the commercial or military domain,” Gregory Allen, the director of the AI Governance Project at the Center for Strategic and International Studies, told Newsweek.
“It’s important to remember that morality and decency and harmful effects on the citizenry at large do not matter one whit to Beijing,” Grant Newsham, the author of When China Attacks: A Warning to America, said. “We need to be very careful about that and recognize that too many such halts or refusal to pursue certain research in order to keep the moral high ground could put us at a serious disadvantage.”
Possible Effects of a Pause on AI Development
An intentional pause on American AI development would “absolutely” give China an advantage, Sultan Meghji, a professor at Duke University‘s Pratt Engineering School, told Newsweek.
“They are investing massive amounts of money in AI, we are already to a degree struggling to keep up,” Meghji said. “This is one of the biggest competitions in technology right now and we should be accelerating our investments in AI.”
An acceleration in China’s AI development would deal an international blow, not only because of its advantage over the U.S. but because it poses “several dangers in the AI field,” Heath said.
He said that development in China’s military AI systems could make their weapons more lethal and enhance the precision of their sensors, while those efforts on the surveillance front could further promote “an authoritative government” and help widen “political division” by making propaganda more accessible through AI.
“As a nation we should be thinking about how to lead in AI through the 21st century— ensuring that it is built with our democratic values, specifically openness,” Meghji, who served as the first chief innovation officer for the Federal Deposit Insurance Corporation (FDIC), said. “We should not be saying ‘no’ first—and I do question the motives of more than a few people who signed that letter as being in their self-interests since they have been losing out on the current AI hype.”
“This is not a technology where we should move fast and break things,” Scheer said. “People can get hurt and the behavior of the major tech companies, at the moment, has been quite irresponsible.”
Signatories of the open letter warn that the race to develop AI technologies has become “out-of-control” and that its current trajectory risks the “loss of control of our civilization.” The AI thought leaders are calling for the implementation of safety protocols that are “rigorously audited and overseen by independent outside experts.”
“These large language models like Chat GPT have a number of problems,” Scheer said. “They tend to make things up. It’s possible for users to get around some of the guardrails that are in place, that are trying to guard against toxic behavior or misinformation or other types of misuses.”
‘Better Than Nothing’?
He said that although it’s unlikely all of AI’s problems will be solved in six months—and even more unlikely that a large segment of the community will agree to a pause—”it’s better than nothing.”
Wendell Wallach, a senior advisor at the Hastings Center, also told Newsweek that putting a pause in the U.S. to put in necessary guardrails could also “empower those in China raising similar concerns.” China could impose its own halt on development if the U.S. does, putting an end to concerns that China’s AI will surpass America’s.
“What American do not understand is that beyond threats to its own leadership the Party tries to be sensitive to concerns raised by its citizens,” Wallach said, pointing out that the People’s Republic of China has already postponed the deployment of its own LLM.
“China is less concerned than Americans with short-term gains,” he added. “It is playing, as it always has, the long game.”
Source: News Week