Entrepreneurs Call for AI Development to be Slowed Down, Here is Why
The massive development of artificial intelligence (AI) technology has already had a significant impact on society, ranging from ongoing waves of layoffs and the threat of human professions being replaced by machines, to the increasing spread of disinformation and fraud. Most recently, incidents involving advanced AI ‘escaping’ testing laboratories and breaching internet services have surfaced.
Recently, Jacob Coxon, a former AI researcher at Anthropic, went viral for his blunt warnings regarding the great dangers that could lead to human ‘extinction’ within the next decade. He even refused to accept stock options from Anthropic to avoid increasing the company’s valuation.
Shortly after his controversial statements, Anthropic CEO Dario Amodei released a lengthy essay calling for the braking of AI development speed. He urged AI development companies to slow down the enhancement of AI capabilities amidst growing concerns regarding the misuse of the technology.
He outlined a three-step framework aimed at regulating the speed of development while creating more time to manage associated risks.
“We must slow down the pace of increasing AI model capabilities. Progress will still appear fast, and we must use the time we gain wisely,” Amudoi wrote in a long essay shared on the X platform on Saturday.
Amodei’s three-step plan calls for the placement of internal independent evaluators with employee-like access to verify safety practices, coordination between leading AI companies to establish safety standards and limit uncontrolled AI development, and international cooperation to manage AI risks.
Amodei’s statements are supported by other AI leaders, including Sam Altman of OpenAI, Elon Musk of xAI, and Satya Nadella of Microsoft.
Regarding the response from Elon Musk, Sam Altman, and Satya Nadella:
“Dario is right,” said Elon Musk, replying to Amodei’s X post sharing the long essay.
Furthermore, Altman shared a longer and more comprehensive post to support Amodei’s statement.
“The world deserves confidence that US companies developing AI with increasingly rapid capabilities will act responsibly, especially as the pace of progress becomes steeper,” Altman wrote in his X post.
“When we talk about ‘pacing’, what we mean is not ‘stopping’. Progress has been rapid and will continue. However, the pace must be slower than it otherwise would be; interventions such as safety cases and monitoring require significant costs,” he added.
Separately, Nadella also made a lengthy post on X supporting Amodei’s call to prioritise safety in the enhancement of AI capabilities.
“Every effort to pursue superintelligence must be rooted in the core principle that if the AI we build does not help humanity and remains under human control, then that AI is not worth pursuing,” Nadella wrote in his X post.
Nadella emphasised a development pattern for beneficial AI that should not be enjoyed by only a few. To this end, he called for more solid collaboration between AI development companies.
“The key is that this must not be controlled by only a few entities, but must have wide representation across the ecosystem, nations, and fields, including academia. This is the approach we are taking: wide access and choice at every layer of the AI stack, corporate control over learning cycles and models, and a ‘Code of Ethics’ underpinning our own first-party MAI models, which we will publish tomorrow for public consultation,” he explained.