{"id":663391,"date":"2025-04-17T22:23:58","date_gmt":"2025-04-17T19:23:58","guid":{"rendered":"https:\/\/en.buradabiliyorum.com\/googles-latest-ai-model-report-lacks-key-safety-details-experts-say\/"},"modified":"2025-04-17T22:23:58","modified_gmt":"2025-04-17T19:23:58","slug":"googles-latest-ai-model-report-lacks-key-safety-details-experts-say","status":"publish","type":"post","link":"https:\/\/buradabiliyorum.com\/en\/googles-latest-ai-model-report-lacks-key-safety-details-experts-say\/","title":{"rendered":"Google&#8217;s latest AI model report lacks key safety details, experts say"},"content":{"rendered":"<div>\n<p id=\"speakable-summary\" class=\"wp-block-paragraph\">On Thursday, weeks after launching its most powerful AI model yet, Gemini 2.5 Pro, Google published a <a rel=\"nofollow\" target=\"_blank\" href=\"https:\/\/storage.googleapis.com\/model-cards\/documents\/gemini-2.5-pro-preview.pdf\" target=\"_blank\" rel=\"noreferrer noopener nofollow\">technical report<\/a> showing the results of its internal safety evaluations. However, the report is light on the details, experts say, making it difficult to determine which risks the model might pose.<\/p>\n<p class=\"wp-block-paragraph\">Technical reports provide useful \u2014 and unflattering, at times \u2014 info that companies don\u2019t always widely advertise about their AI. By and large, the AI community sees these reports as good-faith efforts to support independent research and safety evaluations.<\/p>\n<p class=\"wp-block-paragraph\">Google takes a different safety reporting <a href=\"https:\/\/buradabiliyorum.com\/en\/category\/download-scripts-themes-apps\/\" data-internallinksmanager029f6b8e52c=\"9\" title=\"Download Scripts &amp; Themes &amp; Apps\" target=\"_blank\" rel=\"noopener\">app<\/a>roach than some of its AI rivals, publishing technical reports only once it considers a model to have graduated from the \u201cexperimental\u201d stage. The company also doesn\u2019t include findings from all of its \u201cdangerous capability\u201d evaluations in these write-ups; it reserves those for a separate audit.<\/p>\n<p class=\"wp-block-paragraph\">Several experts TechCrunch spoke with were still disappointed by the sparsity of the Gemini 2.5 Pro report, however, which they noted doesn\u2019t mention Google\u2019s <a rel=\"nofollow\" target=\"_blank\" href=\"https:\/\/deepmind.google\/discover\/blog\/introducing-the-frontier-safety-framework\/\" target=\"_blank\" rel=\"noreferrer noopener nofollow\">Frontier Safety Framework (FSF)<\/a>. Google introduced the FSF last year in what it described as an effort to identify future AI capabilities that could cause \u201csevere harm.\u201d<\/p>\n<p class=\"wp-block-paragraph\">\u201cThis [report] is very sparse, contains minimal information, and came out weeks after the model was already made available to the public,\u201d Peter Wildeford, co-founder of the Institute for AI Policy and Strategy, told TechCrunch. \u201cIt\u2019s impossible to verify if Google is living up to its public commitments and thus impossible to assess the safety and security of their models.\u201d<\/p>\n<p class=\"wp-block-paragraph\">Thomas Woodside, co-founder of the Secure AI Project, said that while he\u2019s glad Google released a report for Gemini 2.5 Pro, he\u2019s not convinced of the company\u2019s commitment to delivering timely supplemental safety evaluations. Woodside pointed out that the last time Google published the results of dangerous capability tests was in June 2024 \u2014 for a model announced in February that same year.<\/p>\n<p class=\"wp-block-paragraph\">Not inspiring much confidence, Google hasn\u2019t made available a report for Gemini 2.5 Flash, a smaller, more efficient model the company announced last week. A spokesperson told TechCrunch a report for Flash is \u201ccoming soon.\u201d<\/p>\n<p class=\"wp-block-paragraph\">\u201cI hope this is a promise from Google to start publishing more frequent updates,\u201d Woodside told TechCrunch. \u201cThose updates should include the results of evaluations for models that haven\u2019t been publicly deployed yet, since those models could also pose serious risks.\u201d<\/p>\n<p class=\"wp-block-paragraph\">Google may have been one of the first AI labs to propose standardized reports for models, but it\u2019s not the only one that\u2019s been accused of underdelivering on transparency lately. Meta released a <a rel=\"nofollow\" target=\"_blank\" href=\"https:\/\/ai.meta.com\/blog\/llama-4-multimodal-intelligence\/\" target=\"_blank\" rel=\"noreferrer noopener nofollow\">similarly skimpy safety evaluation<\/a> of its new Llama 4 open models, and OpenAI opted not to publish any report for its GPT-4.1 <a href=\"https:\/\/buradabiliyorum.com\/en\/category\/watch-movies-tv-seriess\/\" data-internallinksmanager029f6b8e52c=\"8\" title=\"Watch Movies &amp; TV Series\" target=\"_blank\" rel=\"noopener\">series<\/a>.<\/p>\n<p class=\"wp-block-paragraph\">Hanging over Google\u2019s head are assurances the tech giant made to regulators to maintain a high standard of AI safety testing and reporting. Two years ago, <a rel=\"nofollow\" target=\"_blank\" href=\"https:\/\/www.transformernews.ai\/p\/google-breaks-its-promises\" target=\"_blank\" rel=\"noreferrer noopener nofollow\">Google told the U.S. government<\/a> it would publish safety reports for all \u201csignificant\u201d public AI models \u201cwithin scope.\u201d The company\u00a0<a rel=\"nofollow\" target=\"_blank\" href=\"https:\/\/www.itpro.com\/technology\/artificial-intelligence\/microsoft-amazon-meta-and-google-just-promised-to-halt-ai-development-if-models-are-too-dangerous-but-will-they-stick-to-their-promise\" target=\"_blank\" rel=\"noreferrer noopener nofollow\">followed up that promise with similar commitments<\/a>\u00a0to\u00a0<a rel=\"nofollow\" target=\"_blank\" href=\"https:\/\/www.gov.uk\/government\/publications\/frontier-ai-safety-commitments-ai-seoul-summit-2024\/frontier-ai-safety-commitments-ai-seoul-summit-2024\" target=\"_blank\" rel=\"noreferrer noopener nofollow\">other countries<\/a>, pledging to \u201cprovide public transparency\u201d around AI products.<\/p>\n<p class=\"wp-block-paragraph\">Kevin Bankston, a senior adviser on AI governance at the Center for Democracy and <a href=\"https:\/\/buradabiliyorum.com\/en\/category\/technology\/\" data-internallinksmanager029f6b8e52c=\"4\" title=\"Technology\" target=\"_blank\" rel=\"noopener\">Technology<\/a>, called the trend of sporadic and vague reports a \u201crace to the bottom\u201d on AI safety.<\/p>\n<p class=\"wp-block-paragraph\">\u201cCombined with reports that competing labs like OpenAI have shaved their safety testing time before release from months to days, this meager documentation for Google\u2019s top AI model tells a troubling story of a race to the bottom on AI safety and transparency as companies rush their models to market,\u201d he told TechCrunch.<\/p>\n<p class=\"wp-block-paragraph\">Google has said in statements that, while not detailed in its technical reports, it conducts safety testing and \u201cadversarial red teaming\u201d for models ahead of release.<\/p>\n<\/div>\n<blockquote><p><strong><span style=\"color: #ff6600;\">If you liked the article, do not forget to share it with your friends. Follow us on\u00a0<span style=\"color: #ff0000;\"><a style=\"color: #ff0000;\" href=\"https:\/\/news.google.com\/publications\/CAAqBwgKMN63nwsw68G3Aw\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Google News<\/a><\/span>\u00a0too, click on the star and choose us from your favorites.<\/span><\/strong><\/p><\/blockquote>\n<blockquote>\n<p style=\"text-align: center;\"><strong>If you want to read more like this article, you can visit our <span style=\"color: #ff9900;\"><a style=\"color: #ff9900;\" href=\"https:\/\/en.buradabiliyorum.com\/category\/technology\/\" target=\"_blank\" >Technology<\/a><\/span> category.<\/strong><\/p>\n<\/blockquote>\n<p><span style=\"color: black;\"><a style=\"color: #ff9900;\" href=\"https:\/\/techcrunch.com\/2025\/04\/17\/googles-latest-ai-model-report-lacks-key-safety-details-experts-say\/\" target=\"_blank\" >Source<\/a><\/span><\/p>\n","protected":false},"excerpt":{"rendered":"<p>On Thursday, weeks after launching its most powerful AI model yet, Gemini 2.5 Pro, Google published a technical report showing the results of its internal safety evaluations. However, the report is light on the details, experts say, making it difficult to determine which risks the model might pose. Technical reports provide useful \u2014 and unflattering,&#8230;<\/p>\n","protected":false},"author":1,"featured_media":663392,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"fifu_image_url":"https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/03\/GettyImages-2169339854.jpg?resize=1200,857","fifu_image_alt":"","footnotes":""},"categories":[18],"tags":[77337,74864,26293],"class_list":["post-663391","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-technology","tag-ai","tag-gemini","tag-google"],"_links":{"self":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/posts\/663391","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/comments?post=663391"}],"version-history":[{"count":0,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/posts\/663391\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/media\/663392"}],"wp:attachment":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/media?parent=663391"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/categories?post=663391"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/tags?post=663391"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}