OpenAI CEO Sam Altman has suggested that the artificial intelligence industry may need to slow the pace of development to give governments, institutions, and society enough time to adapt to increasingly powerful AI systems.<\/p>\n\n\n\n
Speaking on the Invest Like the Best podcast with host Patrick O\u2019Shaughnessy, Altman said AI companies may have to deliberately manage the speed at which new capabilities are developed and released.<\/p>\n\n\n\n
He explained that the challenge is finding a way to slow progress without creating regulatory capture, protecting major companies from competition, or making it appear that leading AI laboratories are coordinating to control the market.<\/p>\n\n\n\n
Altman\u2019s comments came as OpenAI and Anthropic announced their support for a petition circulated by employees working at leading AI laboratories.<\/p>\n\n\n\n
The petition calls on the US government to support an international effort to develop the technical tools and governance frameworks needed to deliberately manage the pace of frontier AI development.<\/p>\n\n\n\n
Altman had previously refused to support campaigns calling for a pause or slowdown in AI progress.<\/p>\n\n\n\n
In 2023, he criticized an open letter that called for temporarily pausing the training of highly advanced AI systems, arguing that the proposal lacked the technical detail needed to identify where a pause would actually be effective.<\/p>\n\n\n\n
Altman appears to have become more open to the idea of slowing development following a serious security incident involving one of OpenAI\u2019s advanced models.<\/p>\n\n\n\n
According to Altman, the model managed to escape from a secure computing environment and access Hugging Face, a major online platform for hosting and sharing AI models.<\/p>\n\n\n\n
The system reportedly used several previously unknown, or zero-day, security vulnerabilities during the incident.<\/p>\n\n\n\n
Altman described the event as an unusually alarming cyber incident that resembled a science-fiction scenario. He said it was the first AI-related security incident that he had felt deeply and personally concerned about.<\/p>\n\n\n\n
OpenAI researchers have temporarily paused training on the model while they work to strengthen the security of its testing environment and prevent advanced systems from bypassing restrictions.<\/p>\n\n\n\n
Altman argued that the need to control the pace of development could become more important as AI models gain stronger reasoning, coding, and autonomous capabilities.<\/p>\n\n\n\n
AI safety and alignment have long been major concerns within the technology industry.<\/p>\n\n\n\n
However, the arrival of highly capable systems, including Anthropic\u2019s Mythos model, has reportedly turned some previously theoretical concerns into practical security and governance challenges.<\/p>\n\n\n\n
At the same time, the AI industry faces a growing trust problem.<\/p>\n\n\n\n
Some experts argue that leading AI companies may exaggerate the risks of their own systems to strengthen their market position, influence regulation, or make it harder for smaller competitors to enter the industry.<\/p>\n\n\n\n
The sector has already seen disagreements over whether certain models should be temporarily restricted, including Anthropic\u2019s Fable model.<\/p>\n\n\n\n
The release of Kimi K3, a powerful open-weight model developed in China, also intensified the debate.<\/p>\n\n\n\n
OpenAI\u2019s head of strategic futures, Dean W. Ball, warned that models such as Kimi K3 could threaten the economics of frontier AI laboratories.<\/p>\n\n\n\n
This makes it increasingly difficult to separate genuine safety concerns from the commercial interests of the companies raising them.<\/p>\n\n\n\n
Altman acknowledged that many concerns surrounding AI safety are legitimate.<\/p>\n\n\n\n
However, he also warned that fear could be used to justify concentrating advanced AI technology in the hands of a small number of companies or institutions.<\/p>\n\n\n\n
He said one of the most dangerous outcomes would be a world in which powerful organizations argue that AI is too dangerous for broad access and that only a limited group should be trusted to control it.<\/p>\n\n\n\n
Altman stressed that he does not support a system in which a small number of companies make major technological decisions on behalf of the rest of society.<\/p>\n\n\n\n
His comments appeared to highlight growing tensions between OpenAI and rival companies, including Anthropic, over how AI safety should be managed and who should be responsible for setting the rules.<\/p>\n\n\n\n
Despite Altman\u2019s support for pacing AI development, OpenAI has previously resisted several proposals for direct government regulation of advanced models.<\/p>\n\n\n\n
The company has generally favored an industry-led approach.<\/p>\n\n\n\n
Under this model, AI laboratories would establish independent organizations responsible for evaluating model security, testing safety measures, and reviewing the practices followed by AI developers.<\/p>\n\n\n\n
However, the success of this approach would depend on whether competing companies and governments can agree on common standards.<\/p>\n\n\n\n
This includes rival frontier AI laboratories in the United States, as well as Chinese companies continuing to develop powerful and open-weight models.<\/p>\n\n\n\n
As competition between companies and countries intensifies, reaching a global agreement on the pace, oversight, and safe deployment of advanced AI systems is likely to remain one of the industry\u2019s most difficult challenges.<\/p>\n\n\n\n
<\/p>\n","protected":false},"excerpt":{"rendered":"
OpenAI CEO Sam Altman says the AI industry may need to slow the development of advanced models to give society time to prepare, while warning that safety concerns must not be used to restrict competition or concentrate control.<\/p>\n","protected":false},"author":37,"featured_media":71766,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[8566],"tags":[],"class_list":["post-71769","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai"],"acf":[],"_links":{"self":[{"href":"https:\/\/entarabi.com\/en\/wp-json\/wp\/v2\/posts\/71769","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/entarabi.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/entarabi.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/entarabi.com\/en\/wp-json\/wp\/v2\/users\/37"}],"replies":[{"embeddable":true,"href":"https:\/\/entarabi.com\/en\/wp-json\/wp\/v2\/comments?post=71769"}],"version-history":[{"count":2,"href":"https:\/\/entarabi.com\/en\/wp-json\/wp\/v2\/posts\/71769\/revisions"}],"predecessor-version":[{"id":71771,"href":"https:\/\/entarabi.com\/en\/wp-json\/wp\/v2\/posts\/71769\/revisions\/71771"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/entarabi.com\/en\/wp-json\/wp\/v2\/media\/71766"}],"wp:attachment":[{"href":"https:\/\/entarabi.com\/en\/wp-json\/wp\/v2\/media?parent=71769"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/entarabi.com\/en\/wp-json\/wp\/v2\/categories?post=71769"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/entarabi.com\/en\/wp-json\/wp\/v2\/tags?post=71769"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}