{"id":5911,"date":"2025-08-27T23:56:45","date_gmt":"2025-08-27T23:56:45","guid":{"rendered":"https:\/\/jobuzo.com\/en\/openai-and-anthropic-conducted-safety-evaluations-of-each-others-ai-systems\/"},"modified":"2025-08-27T23:56:45","modified_gmt":"2025-08-27T23:56:45","slug":"openai-and-anthropic-conducted-safety-evaluations-of-each-others-ai-systems","status":"publish","type":"post","link":"https:\/\/jobuzo.com\/en\/openai-and-anthropic-conducted-safety-evaluations-of-each-others-ai-systems\/","title":{"rendered":"OpenAI and Anthropic conducted safety evaluations of each other&#8217;s AI systems"},"content":{"rendered":"<div>\n<div><\/div>\n<p class=\"col-body mb-4 leading-7 text-[18px] md:leading-8 break-words min-w-0 engadget-charcoal\">Most of the time, AI companies are locked in a race to the top, treating each other as rivals and competitors. Today, OpenAI and Anthropic revealed that they agreed to evaluate the alignment of each other&rsquo;s publicly available systems and shared the results of their analyses. The full reports get pretty technical, but are worth a read for anyone who&rsquo;s following the nuts and bolts of AI development. A broad summary showed some flaws with each company&rsquo;s offerings, as well as revealing pointers for how to improve future safety tests.<\/p>\n<p class=\"col-body mb-4 leading-7 text-[18px] md:leading-8 break-words min-w-0 engadget-charcoal\">Anthropic said it <ins>evaluated OpenAI models<\/ins> for &ldquo;sycophancy, whistleblowing, self-preservation, and supporting human misuse, as well as capabilities related to undermining AI safety evaluations and oversight.&rdquo; Its review found that o3 and o4-mini models from OpenAI fell in line with results for its own models, but raised concerns about possible misuse with the &#8203;&#8203;GPT-4o and GPT-4.1 general-purpose models. The company also said sycophancy was an issue to some degree with all tested models except for o3.<\/p>\n<p class=\"col-body mb-4 leading-7 text-[18px] md:leading-8 break-words min-w-0 engadget-charcoal\">Anthropic&rsquo;s tests did not include OpenAI&rsquo;s most recent release. <ins>GPT-5<\/ins> has a feature called Safe Completions, which is meant to protect users and the public against potentially dangerous queries. OpenAI recently faced its <ins>first wrongful death lawsuit<\/ins> after a tragic case where a teenager discussed attempts and plans for suicide with ChatGPT for months before taking his own life.<\/p>\n<div class=\"col-fullbleed mb-4 bg-marshmallow pb-5 dark:bg-ramones md:invisible md:mb-0 md:h-0 md:overflow-hidden md:pb-0\">\n<p>Advertisement<\/p>\n<div class=\"flex w-full flex-nowrap justify-center\">\n<div class=\"flex\" id=\"_R3bljlr4kutrilbH1_\">\n<p>Advertisement<\/p>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"col-body mb-4 hidden pb-5 dark:bg-ramones md:block\">\n<div class=\"flex w-full flex-nowrap justify-center\">\n<div class=\"flex\" id=\"_R5bljlr4kutrilbH1_\">\n<p>Advertisement<\/p>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"internal-linking-related-contents\"><a href=\"https:\/\/jobuzo.com\/en\/12-weeks-jail-for-school-it-support-technician-who-took-upskirt-videos-of-teachers\/\" class=\"template-1\"><span class=\"cta\">News :<\/span><span class=\"postTitle\">&lt;div&gt;12 weeks' jail for school IT support technician who took upskirt videos of teachers&lt;\/div&gt;<\/span><\/a><\/div><p class=\"col-body mb-4 leading-7 text-[18px] md:leading-8 break-words min-w-0 engadget-charcoal\">On the flip side, OpenAI <ins>ran tests on Anthropic models<\/ins> for instruction hierarchy, jailbreaking, hallucinations and scheming. The Claude models generally performed well in instruction hierarchy tests, and had a high refusal rate in hallucination tests, meaning they were less likely to offer answers in cases where uncertainty meant their responses could be wrong.<\/p>\n<p class=\"col-body mb-4 leading-7 text-[18px] md:leading-8 break-words min-w-0 engadget-charcoal\">The move for these companies to conduct a joint assessment is intriguing, particularly since OpenAI allegedly violated Anthropic&rsquo;s terms of service by having programmers use Claude in the process of building new GPT models, which led to Anthropic <ins>barring<\/ins> OpenAI&rsquo;s access to its tools earlier this month. But safety with AI tools has become a bigger issue as more critics and legal experts seek guidelines to protect users, particularly minors.<\/p>\n<\/div>\n<p><sub><\/sub><\/p>\n<div>OpenAI and Anthropic conducted safety evaluations of each other&rsquo;s AI systems<\/div>\n<p><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Most of the time, AI companies are locked in a race to the top, treating each other as rivals and competitors. Today, OpenAI and Anthropic revealed that they agreed to evaluate the alignment of each other&rsquo;s publicly available systems and shared the results of their analyses. The full reports get pretty technical, but are worth&#8230;<\/p>\n<p class=\"more-link-wrap\"><a href=\"https:\/\/jobuzo.com\/en\/openai-and-anthropic-conducted-safety-evaluations-of-each-others-ai-systems\/\" class=\"more-link\">Read More<span class=\"screen-reader-text\"> &ldquo;OpenAI and Anthropic conducted safety evaluations of each other&#8217;s AI systems&rdquo;<\/span> &raquo;<\/a><\/p>\n","protected":false},"author":1,"featured_media":5912,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-5911","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-news"],"_links":{"self":[{"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/posts\/5911","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/comments?post=5911"}],"version-history":[{"count":0,"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/posts\/5911\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/media\/5912"}],"wp:attachment":[{"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/media?parent=5911"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/categories?post=5911"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/tags?post=5911"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}