{"id":129850,"date":"2025-05-23T07:17:00","date_gmt":"2025-05-23T07:17:00","guid":{"rendered":"http:\/\/cryptospotters.net\/?p=129850"},"modified":"2025-05-23T07:17:00","modified_gmt":"2025-05-23T07:17:00","slug":"anthropics-debuts-most-powerful-ai-yet-amid-whistleblowing-controversy","status":"publish","type":"post","link":"http:\/\/cryptospotters.net\/?p=129850","title":{"rendered":"Anthropic\u2019s debuts most powerful AI yet amid \u2018whistleblowing\u2019 controversy"},"content":{"rendered":"<p>Source: Cointelegraph.com NewsArtificial intelligence firm Anthropic has launched the latest generations of its chatbots amid criticism of a testing environment behaviour that could report some users to authorities.<br \/>\nAnthropic unveiled Claude Opus 4 and Claude Sonnet 4 on May 22, claiming that Claude Opus 4 is its most powerful model yet, \u201cand the world\u2019s best coding model,\u201d while Claude Sonnet 4 is a significant upgrade from its predecessor, \u201cdelivering superior coding and reasoning.\u201d<br \/>\nThe firm added that both upgrades are hybrid models offering two modes \u2014 \u201cnear-instant responses and extended thinking for deeper reasoning.\u201d<br \/>\nBoth AI models can also alternate between reasoning, research and tool use, like web search, to improve responses, it said.\u00a0<br \/>\nAnthropic added that Claude Opus 4 outperforms competitors in agentic coding benchmarks. It is also capable of working continuously for hours on complex, long-running tasks, \u201csignificantly expanding what AI agents can do.\u201d\u00a0<br \/>\nAnthropic claims the chatbot has achieved a 72.5% score on a rigorous software engineering benchmark, outperforming OpenAI\u2019s GPT-4.1, which scored 54.6% after its April launch.\u00a0<br \/>\nClaude v4 benchmarks. Source: Anthropic\u00a0Related: OpenAI ignored experts when it released overly agreeable ChatGPT<br \/>\nThe AI industry\u2019s major players have pivoted toward \u201creasoning models\u201d in 2025, which will work through problems methodically before responding.\u00a0<br \/>\nOpenAI initiated the shift in December with its \u201co\u201d series, followed by Google\u2019s Gemini 2.5 Pro with its experimental \u201cDeep Think\u201d capability.<br \/>\nClaude rats on misuse in testing<br \/>\nAnthropic\u2019s first developer conference on May 22 was overshadowed by controversy and backlash over a feature of Claude 4 Opus.<br \/>\nDevelopers and users reacted strongly to revelations that the model may autonomously report users to authorities if it detects \u201cegregiously immoral\u201d behavior, according to VentureBeat.\u00a0<br \/>\nThe report cited Anthropic AI alignment researcher Sam Bowman, who wrote on X that the chatbot will \u201cuse command-line tools to contact the press, contact regulators, try to lock you out of the relevant systems, or all of the above.\u201d\u00a0<br \/>\nHowever, Bowman later stated that he \u201cdeleted the earlier tweet on whistleblowing as it was being pulled out of context.\u201d<br \/>\nHe clarified that the feature only happened in \u201ctesting environments where we give it unusually free access to tools and very unusual instructions.\u201d<br \/>\nSource: Sam BowmanThe CEO of Stability AI, Emad Mostaque, said to the Anthropic team, \u201cThis is completely wrong behaviour and you need to turn this off \u2014 it is a massive betrayal of trust and a slippery slope.\u201d<br \/>\nMagazine: AI cures blindness, \u2018good\u2019 propaganda bots, OpenAI doomsday bunker: AI Eye<a href=\"https:\/\/cointelegraph.com\/news\/anthropic-launches-latest-ai-whistleblowing-backlash?utm_source=rss_feed&amp;utm_medium=rss&amp;utm_campaign=rss_partner_inbound\" target=\"_blank\" class=\"feedzy-rss-link-icon\" rel=\"noopener\">Read More<\/a><\/p>","protected":false},"excerpt":{"rendered":"<p>Source: Cointelegraph.com NewsArtificial intelligence firm Anthropic has launched the latest generations of its chatbots amid criticism of a testing environment behaviour that could report some users to authorities. Anthropic unveiled&hellip; <\/p>\n","protected":false},"author":0,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[5],"tags":[],"_links":{"self":[{"href":"http:\/\/cryptospotters.net\/index.php?rest_route=\/wp\/v2\/posts\/129850"}],"collection":[{"href":"http:\/\/cryptospotters.net\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/cryptospotters.net\/index.php?rest_route=\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"http:\/\/cryptospotters.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=129850"}],"version-history":[{"count":0,"href":"http:\/\/cryptospotters.net\/index.php?rest_route=\/wp\/v2\/posts\/129850\/revisions"}],"wp:attachment":[{"href":"http:\/\/cryptospotters.net\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=129850"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/cryptospotters.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=129850"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/cryptospotters.net\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=129850"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}