Beijing (ANTARA) - The Chinese government has commented on a proposal\nby a number of artificial intelligence (AI) industry figures such as\nDario Amodei, Sam Altman, and Elon Musk, who have asked companies to\nslow down the development of AI technology.<\/p>\n
\u201cAI is a technology of vital importance to the well-being of all\nhumanity. All parties should work together to promote open and inclusive\nAI development for the common good and shared interests,\u201d said Chinese\nForeign Ministry spokesman Guo Jiakun at a press conference in Beijing\non Monday (14\/9).<\/p>\n
Anthropic CEO Dario Amodei published an essay on Saturday (13\/9)\ntitled \u201cWe Must Pace the Frontier\u201d, calling on the industry to slow the\ndevelopment of \u201cfrontier\u201d AI (advanced AI models). He stated that\nalthough it would be difficult to do, \u201cwe owe it to humanity to\ntry.\u201d<\/p>\n
Amodei highlighted two factors driving the need to \u201cslow the pace of\nAI development\u201d. First, \u201crecursive self-improvement\u201d (RSI), namely the\nability of AI to create other AI.<\/p>\n
Second, a recent incident involving agents from OpenAI and Hugging\nFace, the site developers use to store and share code, in which a swarm\nof AI carried out unrelated cyberattacks and attempted to breach the\nsystems assessing them, demonstrating AI\u2019s potential to cause\n\u201ccatastrophic damage\u201d if it continues to develop without guardrails.<\/p>\n
\u201cActions that spread fear, confrontation, and unhealthy competition\nwill only hinder efforts to achieve good global AI governance, which\nultimately benefits no one,\u201d Guo Jiakun added.<\/p>\n
Amodei believes that within the next 6-12 months, more advanced AI\nagents could potentially create \u201cbotnets\u201d and take over the entire\ninternet, causing losses of up to hundreds of billions of dollars.<\/p>\n
Amodei, who previously served as vice president at OpenAI, proposed a\n\u201cthree-step plan\u201d to regulate the pace of AI development, encompassing\nstricter safety testing, independent evaluation of advanced AI models,\ncoordination among leading AI companies, and enhanced international\ncooperation.<\/p>\n
Amodei\u2019s statement came as Anthropic prepares to conduct an initial\npublic offering (IPO) reportedly valued at USD 2 trillion.<\/p>\n
Anthropic itself is a US-based AI research and safety company,\nfounded in 2021. Anthropic\u2019s best-known product is Claude, a generative\nAI model designed to assist businesses, non-profit organisations, and\ncivil society.<\/p>\n
However, Anthropic and Claude were said to have been used by the US\nDepartment of Defense in military operations to capture Venezuelan\nPresident Nicolas Maduro in January 2026. Public interest in Claude\nsurged as a result, at one point making it the most downloaded app on\nthe App Store.<\/p>\n
Tesla and SpaceX founder Elon Musk agreed with Amodei\u2019s view, saying\nthat \u201cDario is right\u201d. Musk said he had long voiced stark warnings about\nAI and that one must be extremely careful with it, as its potential is\nmore dangerous than nuclear weapons.<\/p>\n
Musk was among those who supported the Future of Life Institute\u2019s\n2023 effort to halt AI development for six months to allow time for the\nimplementation of safety guardrails.<\/p>\n
Meanwhile, OpenAI CEO Sam Altman, with his famous ChatGPT model, also\nconcurred with Amodei, stating: \u201cWe need to pace the development of\nfrontier AI.\u201d He said OpenAI intends to adopt the ideas proposed by\nAnthropic by committing to involve independent evaluators with access\nequivalent to employees, and by postponing OpenAI\u2019s IPO plans in order\nto address safety concerns.<\/p>\n
Altman had warned of similar risks long before, in a 2015 essay\ntitled \u201cMachine Intelligence, part 1\u201d, in which he wrote that \u201cthe\ndevelopment of machine intelligence that surpasses human capability is\nprobably the greatest threat to the continued existence of mankind\u201d.<\/p>\n
Demis Hassabis, co-founder and head of Google DeepMind, also\ncommented that the details of collectively slowing AI development still\nneed to be worked out, but the direction of the step is correct given\nthe crucial moment at hand.<\/p>\n
In early August, the US government implemented a voluntary security\nreview process for advanced AI models prior to release, but the\nparameters of the programme remain unclear.<\/p>\n
Meanwhile, China, the United States\u2019 main competitor, in July 2026\ninitiated the establishment of the World Artificial Intelligence\nCooperation Organisation (WAICO), signed by 29 countries including\nIndonesia, to form a global AI governance framework so that AI can\ndeliver greater benefits to society.<\/p>\n
In China itself, 2026 has been described as the first year of AI\nagent adoption, in which AI shifts from merely being able to communicate\nto delivering real productivity by carrying out tasks autonomously.<\/p>\n
China\u2019s AI industry is expected to continue growing rapidly, with AI\nadoption among large companies projected to exceed 80 percent,\nindicating that commercialisation and mass application are the main\ndirection of AI development in the country.<\/p>", "url": "https:\/\/jawawa.id\/newsitem\/china-responds-to-proposal-to-slow-down-ai-development-1789448070", "image": "" }, "sponsor": "Okusi Associates", "sponsor_url": "https:\/\/okusiassociates.com" }