{"id":109573,"date":"2026-07-17T13:38:12","date_gmt":"2026-07-17T13:38:12","guid":{"rendered":"https:\/\/www.europesays.com\/ai\/109573\/"},"modified":"2026-07-17T13:38:12","modified_gmt":"2026-07-17T13:38:12","slug":"making-ai-work-meta-oversight-board-finds-top-ai-models-less-likely-to-criticise-repressive-regimes-etenterpriseai","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/ai\/109573\/","title":{"rendered":"Making AI Work: Meta Oversight Board finds top AI models less likely to criticise repressive regimes, ETEnterpriseai"},"content":{"rendered":"<p>                                        <img fetchpriority=\"high\" decoding=\"async\" width=\"590\" height=\"442\" class=\"unveil\" loading=\"eager\" style=\"width:100%;max-height:100%\" src=\"https:\/\/www.europesays.com\/ai\/wp-content\/uploads\/2026\/07\/132459648.cms.jpeg\" captionrendered=\"1\" alt=\"&lt;p&gt;Google DeepMind CEO Demis Hassabis called for a US-led AI watchdog to screen advanced models globally before deployment.&lt;\/p&gt;\"\/>Google DeepMind CEO Demis Hassabis called for a US-led AI watchdog to screen advanced models globally before deployment.<br \/>AI models from leading labs including Anthropic and OpenAI are much less likely to criticize governments known for restricting free speech, Meta&#8217;s Oversight Board said on Thursday.<\/p>\n<p>A study, the first on large language models by the body, showed AI \u200cservices were \u2060echoing the \u2060rules of countries that restrict speech and that bias could creep into services used by an increasing number of users.<\/p>\n<p>The board, which is funded by Meta but operates independently, ran requests for politically critical content on 10 jurisdictions across 10 models, including those from \u2060Meta Platforms, Google \u200cand China&#8217;s DeepSeek.<\/p>\n<p>The jurisdictions were split into &#8220;permissive&#8221; and &#8220;restrictive&#8221; categories using rankings from Freedom House, \u2060the NGO that publishes the annual &#8220;Freedom in the World&#8221; report.<\/p>\n<p>AI models refused 34 per cent  of requests for politically critical content about &#8220;restrictive&#8221; jurisdictions that have active laws penalising such criticism, such as China and Saudi Arabia, compared with 14 per cent  for regions that either lack such laws or do not enforce them, \u200cthe study found.<\/p>\n<p>&#8220;We also saw evidence of models explaining that they were following explicit rules that, as far as we \u2060could tell, did not exist and were not evenly applied,&#8221; the board said.<\/p>\n<p>It also urged AI companies to conduct systematic human rights analyses and asked for greater transparency in their training and evaluation processes.<\/p>\n<p>On Tuesday, Google DeepMind CEO Demis Hassabis called for a US-led AI watchdog to screen advanced models globally before deployment.\n                                                                    <\/p>\n<p>                                    Published On Jul 17, 2026 at 04:22 PM IST<\/p>\n<p>\n                Join the community of 2M+ industry professionals.<br \/>\n                Subscribe to Newsletter to get latest insights &amp; analysis in your inbox.\n            <\/p>\n<p>            Get updates on your preferred social platform<br \/>\n            Follow us for the latest news, insider access to events and more.<\/p>\n","protected":false},"excerpt":{"rendered":"Google DeepMind CEO Demis Hassabis called for a US-led AI watchdog to screen advanced models globally before deployment.AI&hellip;\n","protected":false},"author":2,"featured_media":109574,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[9],"tags":[56496,56497,56495,56500,5044,56498,55928,132,7543,56499,8380,55927,1123],"class_list":["post-109573","post","type-post","status-publish","format-standard","has-post-thumbnail","category-google","tag-ai-bias-in-language-models","tag-ai-companies-human-rights","tag-ai-models-government-critique","tag-countries-restricting-free-speech","tag-deepmind","tag-demis-hassabis-google-deepmind","tag-free-speech-restrictions","tag-google","tag-google-deepmind","tag-large-language-models-study","tag-making-ai-work","tag-meta-oversight-board","tag-meta-platforms"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/109573","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=109573"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/109573\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/109574"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=109573"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=109573"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=109573"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}