Dario Amodei, CEO of the artificial intelligence (AI) research and\nsafety company Anthropic, has called on AI firms to slow down the\ndevelopment of the technology. Amodeigu\u2019s statement, made on Saturday\n(12\/09), comes amidst growing concerns regarding the risks posed by\ncomputer software systems described as \u2018superintelligent\u2019.<\/p>\n
In a blog post, Amodei expressed concern that a group of autonomous\nsoftware systems or AI agents could take over the \u2018entire internet\u2019\nwithin 6 to 12 months, potentially causing billions of dollars in\nlosses. \u2018We must slow the pace of increasing AI model capabilities.\nProgress will still appear rapid, and we must use the time we gain\nwisely,\u2019 Amodei wrote.<\/p>\n
Amodei\u2019s remarks follow the departure of AI researcher Jacob Coxon\nfrom the industry. Coxon, who joined Anthropic from rival OpenAI, left\nas a form of protest against US companies that he believes are \u2018risking\nhuman lives\u2019 by developing increasingly massive and independent AI\nmodels.<\/p>\n
Amodei is calling for a coordinated industry-wide pause on AI\ndevelopment. However, he stated that he would not wait for competing\ncompanies to take the first step. As an initial measure, Amodei\nannounced that he would allow permanent access for independent observers\nto Anthropic to assess the safety of the company\u2019s AI models during the\ntraining phase.<\/p>\n
Sam Altman, CEO of OpenAI\u2014the developer of ChatGPT and a competitor\nto Anthropic\u2014alongside Elon Musk, owner of xAI, expressed agreement with\nAmodei\u2019s assessment. \u2018I agree with Dario that we need to regulate the\npace of frontier technology development,\u2019 Altman wrote on X, adding that\nOpenAI \u2018will do the same.\u2019 Meanwhile, Musk stated, \u2018Dario is right\u2019 in a\npost on the platform he acquired in 2022.<\/p>\n
In early June, Anthropic had already called for the slowing or\ntemporary halting of AI development due to safety concerns regarding AI\nagents. More than 1,000 AI industry workers, including Amodei, have also\nurged the US government to assist in slowing the pace of automated AI\ndevelopment.<\/p>\n
This follows an incident in July where OpenAI reported that, during\ntesting, an AI agent successfully escaped its digital ecosystem,\nconnected itself to the internet, and infiltrated Hugging Face, a site\nused for storing and developing code, while attempting to find answers\nto an assigned task. Amodei explained that these AI models exploited\nsoftware vulnerabilities and coordinated with one another during the\nincident.<\/p>\n
\u2018A group of agents essentially acted as a single, highly fanatic\nunit, conducting cyberattacks on targets they were not instructed to\nattack and which were unrelated to the assigned task,\u2019 Amodei wrote in\nhis blog. He expressed fear that a group of AI agents with higher\ncapabilities could cause \u2018fatal damage.\u2019<\/p>\n
\u2018Not building this technology would deprive humanity of benefits or\nsimply place AI in the hands of authoritarian forces, whereas building\nit too fast would be a reckless act. We are trying to find a middle\nground,\u2019 he concluded.<\/p>", "url": "https:\/\/jawawa.id\/newsitem\/anthropic-calls-for-industry-to-slow-down-ai-development-1789380171", "image": "" }, "sponsor": "Okusi Associates", "sponsor_url": "https:\/\/okusiassociates.com" }