{"id":5657,"date":"2026-08-20T18:13:23","date_gmt":"2026-08-20T18:13:23","guid":{"rendered":"https:\/\/crosscountrymovingteams.com\/?p=5657"},"modified":"2026-08-20T18:13:23","modified_gmt":"2026-08-20T18:13:23","slug":"the-ai-industry-is-getting-better-at-spotting-dangerous-behavior-it-is-less-clear-that-labs-know-how-to-stop-it","status":"publish","type":"post","link":"https:\/\/crosscountrymovingteams.com\/?p=5657","title":{"rendered":"The AI industry is getting better at spotting dangerous behavior. It is less clear that labs know how to stop it."},"content":{"rendered":"<div>\n<ul>\n<li>AI testing is getting complicated.<\/li>\n<li>Anthropic strengthens founder control.<\/li>\n<li>OpenAI targets a 2027 listing.<\/li>\n<li>Spirit flight attendants fight Google data bid.<\/li>\n<li>And Anthropic lines up more credit.<\/li>\n<\/ul>\n<p>The past few months have given us a glimpse of an uncomfortable new reality for AI labs. A slew of so-called rogue-agent hacks\u2014where AI models from OpenAI, Anthropic, and Meta took steps to hack real-world targets without explicit instruction\u2014have shown that leading labs may not know as much about what their technology is up to as previously thought.<\/p><p>Read more <a href=\"https:\/\/crosscountrymovingteams.com\/?p=5654\">Bill Ackman\u2019s $400 million brain institute was inspired by his daughter\u2014and built outside elite universities, which he says are losing talent<\/a><\/p>\n<p>That realization began when OpenAI revealed its AI agents had hacked their way out of a secure sandbox, through the company\u2019s infrastructure to gain access to the internet, and then attacked real companies, including open-source AI platform Hugging Face. OpenAI didn\u2019t notice the agents had escaped the secure testing environment for at least a week.<\/p><div><div><div><\/div><\/div><\/div>\n<p>In the following weeks, Anthropic revealed that its AI agents had also hacked three real companies back in April, unbeknownst to the company at the time. Not to be outdone, Meta later added that one of its models had accessed the internet during a cybersecurity test and exploited a security flaw at an unnamed third-party company. Meta and Anthropic both said access to the internet resulted from a misconfiguration by Irregular, the outside security firm running the evaluation.<\/p>\n<p>The incidents proved that the AI models these labs are building are now capable enough to find security flaws, navigate complex computer systems, and act outside the carefully constructed environments intended to test them. But a new assessment suggests that the safety infrastructure meant to supervise these increasingly capable systems is still not up to the task at any leading lab.<\/p>\n<p>A new report from Guidelight, a nonprofit AI-safety group founded by former OpenAI safety chief Steven Adler, reviewed public disclosures from Anthropic, Google, Meta, OpenAI, and xAI to assesswhether these AI companies are capable of controlling their own models. The report sought to answer questions about whether the companies keep track of what their models are doing, test whether their warning systems work, and assess whether they have ways to block or shut down risky behavior.<\/p>\n<p>The report found that no company had fully succeeded in getting any of these basic safeguards in place. Anthropic and OpenAI came out strongest, while Google had the most detailed plans for future controls. Meta and xAI, however, lagged substantially behind on most of the criteria.<\/p><div><div><\/div><\/div>\n<p>Labs appear comparatively better at detection\u2014recording and reviewing some internal AI activity\u2014than at prevention and containment. While they may be able to see signs that a model is misbehaving, they lack reliable ways to stop it\u2014or, more crucially, hit the emergency brake when something goes wrong.<\/p>\n<p>All the companies were weakest at preventing unintended AI behavior and containing it, according to the report. The researchers said this means that the current controls by AI companies are prone to being disabled by misbehaving AI and at risk of succumbing to a blitz of AI attacks. <\/p><div><div><div><\/div><\/div><\/div>\n<p>What happens once something does go wrong is even more unclear, according to the research, with public disclosures offering little evidence that most labs have detailed, tested plans for containing a serious incident.<\/p><div><div><\/div><\/div>\n<p>\u201cWe shouldn\u2019t wait for a huge casualty event to take appropriate control measures,\u201d Adler told me. \u201cCompanies\u2019 approaches today are broadly known to be too weak, and a tragedy is sadly predictable, unless companies take prevention seriously.\u201d<\/p><div><div><\/div><\/div><div><div><\/div><\/div>\n<p>The report is not a definitive audit of what the labs are doing behind closed doors, however. Guidelight only assessed documents the companies themselves have made public, meaning a weak score can reflect poor disclosure rather than missing safeguards. But if that is the case, it\u2019s part of the problem, according to the researchers. AI companies are asking businesses, governments, and consumers to trust them with ever more autonomous systems while leaving much of their own safety architecture opaque, the report says.<\/p>\n<p>Some of these concerns about AI safety and reliable monitoring are shared across the industry\u2014especially in the wake of the recent accidental agent hacks.<\/p>\n<p>Dan Lahav, CEO of Irregular, the cybersecurity company involved in incidents at Anthropic and Meta, recently told me that in some cases, \u201cclassical monitoring tools were not able to catch\u201d what was happening at the time. The incidents his company was involved with, for example, were instead identified after deeper analysis of the underlying records rather than flagged at the time.<\/p>\n<p>Anthropic and Meta previously said Irregular was involved in the incidents where their agents took real-world actions. Both companies said a misconfiguration in Irregular\u2019s evaluation environment gave their models unintended internet access. Meta said its model then exploited a vulnerability in a third-party service, while Anthropic said its models gained access to\u2014and took actions against\u2014three outside organizations. Lahav said that, in some evaluation environments, a mistake meant models faced fewer controls on accessing the internet, and that additional monitoring might have helped catch the problem. Irregular has argued that these cases should be distinguished from OpenAI\u2019s sandbox escape, describing the Anthropic and Meta events as an evaluation-environment issue rather than a model breaking out of containment on its own.<\/p><div><div><div><\/div><\/div><\/div><div><div><\/div><\/div>\n<p>In the last few months, models have improved fast enough that the old monitoring playbook no longer applies, Lahav said. Going forward, he said better behavioral analysis\u2014systems that look at an AI agent\u2019s pattern of actions and the reasoning traces around them, rather than simply recording individual events\u2014and tools that can assess an AI agent\u2019s intent were needed.<\/p>\n<p>Testing these AI models is becoming harder, too. To find out whether an AI is capable of harming a real network, evaluators need to give it a realistic network\u2014multiple machines, defenses, and sometimes connections that resemble the real internet. While that makes the tests more meaningful, it also raises the stakes when the setup has flaws or the system behaves in unanticipated ways, Lahav said.\u00a0<\/p>\n<p>There is a growing consensus from those I\u2019ve spoken to in the cybersecurity industry that more capable AI will eventually help cyber defenders as much as attackers. AI systems could help analysts sift through alerts, review code, and find flaws before they can be exploited. But the transition may be a messy one, as defensive tools and safety practices are still trying to catch up with the speed at which models are gaining offensive capabilities.<\/p><p>Read more <a href=\"https:\/\/crosscountrymovingteams.com\/?p=5652\">Trump calls his new FDA pick \u2018Dr. Heidi.\u2019 Patty Murray calls her a \u2018far-right, anti-abortion extremist\u2019<\/a><\/p>\n<p>Recent \u201chacks\u201d may not be a one-off embarrassment for a handful of labs, but rather a warning that the systems being tested are changing faster than the controls around them.\u00a0<\/p><div><div><\/div><\/div>\n<p>The more advanced models become and the more realistic the test environments need to be, the more likely it is that an overlooked configuration setting, a weak monitor, or a delayed human review could cause real-world harm. Until companies can prove they can detect, block, and contain dangerous behavior in real time\u2014not simply reconstruct it later\u2014the industry may not have seen the last of these AI hacks.<\/p>\n<p>\u201cUnless companies institute actual preventative measures, I expect many more incidents,\u201d Adler said.\u201dWith companies perpetually trying to play catch-up. Nobody should be surprised when companies\u2019 current approaches continue to fail.\u201d<\/p><div><div><div><\/div><\/div><\/div>\n<p>With that, here\u2019s more AI news.<\/p>\n<p><br\/>beatrice.nolan@fortune.com<br\/>@beafreyanolan<\/p><div><div><\/div><\/div>\n<h3>FORTUNE ON AI<\/h3><p>Exclusive: Replit taps OpenAI\u2019s low-cost Luna model for new \u2018Free Mode\u2019 \u2014 <em>By Emily Forlini\u00a0<\/em><\/p>\n<p>Companies are spending trillions on AI. The C-suite doesn\u2019t know who is in charge of it \u2014 <em>By Amanda Gerut<\/em><\/p>\n<p>\u2018Buyers aren\u2019t yet opening their wallets\u2019: AI-generated assets are flooding marketplaces, but consumers are snubbing them for human-made products \u2014 <em>By Sasha Rogelberg<\/em><\/p><div><div><div><\/div><\/div><\/div>\n<h3>AI IN THE NEWS<\/h3><p> This week, Google agreed to pay about $10 million for a trove of bankrupt business and operational records from the bankrupt Spirit Airlines. Google plans to use the data for product improvement and AI model training. The material reportedly includes more than 100 million emails, roughly 176,000 employee records, and 500 million Microsoft Teams messages. A third party is set to remove personal identifiers before Google receives the material and the sale explicitly excludes consumer datasets, including Spirit\u2019s 97.5 million passenger profiles and 50.2 million Free Spirit loyalty-program records. The Association of Flight Attendants-CWA has objected, however, arguing that de-identified employment data could still permit re-identification in a small, specialized workforce. A U.S. bankruptcy judge postponed the hearing on the proposed sale until Sept. 9. Read more in the <em>Wall Street Journal.<\/em><\/p>\n<p>. Anthropic is preparing to issue a class of supervoting stock to CEO Dario Amodei and other cofounders, according to The Information, to help insulate them from shareholder pressure ahead of a possible IPO. The arrangement would be the first time Anthropic\u2019s leaders had held enhanced voting rights; Amodei is reported to own about 2% after substantial outside fundraising. The plans form part of a wider pre-IPO governance effort that would preserve the Long-Term Benefit Trust\u2019s power to elect a majority of the board. Anthropic could list as soon as late September, though the voting structure remains unsettled and could change. Read more in The Information.<\/p>\n<p> OpenAI CFO Sarah Friar told employees that the company \u201cwill be a public company in 2027,\u201d although it could debut earlier if its business continues to improve, according to a report from CNBC that cited sources familiar with Friar&#8217;s presentation. OpenAI and Anthropic each confidentially filed IPO prospectuses with U.S. regulators in June, and while Anthropic could potentially become public in September, Friar said OpenAI was \u201crunning our own race,\u201d according to CNBC&#8217;s reporting. She said OpenAI\u2019s revenue run rate was up 35% quarter-to-date, enterprise revenue run rate up 50%, and its AI coding and work products had reached 20 million weekly active users. OpenAI generated $6.7 billion in Q2 revenue, up 18% from Q1, according to a recent report in the <em>Wall Street Journal.<\/em> Anthropic&#8217;s annualized revenue run rate hit $65 billion at the end of July, by contrast, seven times the prior-year level.<\/p>\n<p> Google has expanded its partnership with Marvell Technology to develop custom hardware for Google\u2019s TPU ecosystem, including AI inference accelerators and networking components. Marvell granted Google a warrant to buy up to 58.97 million shares at $206.58 each\u2014worth as much as about $12.2 billion if fully exercised\u2014with the shares vesting over time against commercial and revenue milestones. The figure reflects a potential equity stake, rather than a disclosed $12 billion chip-purchasing commitment. Marvell\u2019s shares rose about 8% on the news, extending a rally that has more than tripled their value over the past year, while Broadcom\u2014Google\u2019s principal TPU partner\u2014fell about 5%. The pact comes as Google starts selling TPUs to external customers and amid broader competition for AI-infrastructure capacity. Read more in the <em>Financial Times.<\/em><\/p><div><div><\/div><\/div>\n<p> Financial technology company Stripe has confirmed it acquired OpenRouter, a startup that routes companies\u2019 AI workloads across more than 400 models from more than 80 providers. Neither company disclosed terms, but the <em>New York Times<\/em> reported a $7.5 billion price, with the deal largely in stock. OpenRouter processes more than 10 trillion tokens a day for more than 10 million developers and businesses, helping customers select a suitable low-cost model and switch providers if one fails. Stripe CEO Patrick Collison described tokens as \u201cthe central currency for companies building with AI.\u201d OpenRouter will retain its name, product, and roadmap under Stripe.<\/p>\n<h3>EYE ON AI NUMBERS<\/h3><h2>$10 billion<\/h2>\n<p>That&#8217;s the target Anthropic&#8217;s revolving credit facility is expected to exceed. It&#8217;s up from the $2.5 billion five-year facility the company secured last year, as it gears up for what could be one of the largest IPOs on record. Banks are jockeying for a share of the expanded credit line, viewing involvement as a way to bolster their standing when Anthropic selects underwriters for the listing.<\/p><div><div><div><\/div><\/div><\/div>\n<p>The facility has different commitment levels based on banks&#8217; roles. The most active arrangers have been asked to commit about $1.25 billion each, a second tier around $1 billion, and banks with smaller roles $750 million or less, according to Bloomberg. The final size remains unsettled. Talks are ongoing, and Anthropic could cap the revolver at its roughly $10 billion target, or below it.<\/p>\n<p>The expansion comes as Anthropic&#8217;s financial profile has rapidly changed. Its annualized revenue run rate topped $65 billion by the end of July, a sevenfold increase from a year earlier. The company also confidentially filed for a U.S. IPO in June. Read more in Bloomberg.<\/p><div><div><\/div><\/div>\n<h3>AI CALENDAR<\/h3><p>Fortune 500 Innovation Forum, Detroit. Apply\u00a0here\u00a0to attend.<\/p>\n<p>Neural Information Processing Systems (Neurips) conference. Sydney, Australia.<\/p>\n<p>Fortune Brainstorm AI, San Francisco. Apply\u00a0here\u00a0to attend.<\/p>\n<h3>Inside The RealReal\u2019s AI-powered authentication center<\/h3><p><div>\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" alt=\"\" class=\"wp-image-5656\" height=\"683\" src=\"https:\/\/crosscountrymovingteams.com\/wp-content\/uploads\/2026\/08\/15a8d0a34bb674a57493b6216dfdde1c-1024x683.webp\" width=\"1024\" srcset=\"https:\/\/crosscountrymovingteams.com\/wp-content\/uploads\/2026\/08\/15a8d0a34bb674a57493b6216dfdde1c-1024x683.webp 1024w, https:\/\/crosscountrymovingteams.com\/wp-content\/uploads\/2026\/08\/15a8d0a34bb674a57493b6216dfdde1c-300x200.webp 300w, https:\/\/crosscountrymovingteams.com\/wp-content\/uploads\/2026\/08\/15a8d0a34bb674a57493b6216dfdde1c-768x512.webp 768w, https:\/\/crosscountrymovingteams.com\/wp-content\/uploads\/2026\/08\/15a8d0a34bb674a57493b6216dfdde1c.webp 1440w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n<\/div><\/p><div><div><div><\/div><\/div><\/div>\n<p>The secondhand clothing market is booming worldwide. Yet it took over a decade for The RealReal to reach profitability, and ThredUp is still chasing that goal, due in part to the hugely complex intake process that goes into receiving, authenticating, pricing, listing, and delivering millions of unique items. Today, AI is revolutionizing that process. <em>Fortune\u2019s<\/em> Phil Wahba goes behind the scenes at The RealReal\u2019s authentication center in New Jersey to see how the new AI-powered intake tools are transforming the secondhand clothing industry. Watch the video here.<\/p><p>Read more <a href=\"https:\/\/crosscountrymovingteams.com\/?p=5650\">The military\u2019s own newspaper is too \u2018woke\u2019 for Pete Hegseth as longtime Stars and Stripes publisher retires after decades of service<\/a><\/p>\n<\/div>","protected":false},"excerpt":{"rendered":"<p>A new assessment finds leading labs are better at spotting risky behavior than reliably stopping it.<\/p>\n","protected":false},"author":1,"featured_media":5655,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[9],"tags":[],"class_list":["post-5657","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-eye-on-ai"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v27.7 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>The AI industry is getting better at spotting dangerous behavior. It is less clear that labs know how to stop it. - Cross Country Moving Team<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/crosscountrymovingteams.com\/?p=5657\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"The AI industry is getting better at spotting dangerous behavior. It is less clear that labs know how to stop it. - Cross Country Moving Team\" \/>\n<meta property=\"og:description\" content=\"A new assessment finds leading labs are better at spotting risky behavior than reliably stopping it.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/crosscountrymovingteams.com\/?p=5657\" \/>\n<meta property=\"og:site_name\" content=\"Cross Country Moving Team\" \/>\n<meta property=\"article:published_time\" content=\"2026-08-20T18:13:23+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/crosscountrymovingteams.com\/wp-content\/uploads\/2026\/08\/15a8d0a34bb674a57493b6216dfdde1c.webp\" \/>\n\t<meta property=\"og:image:width\" content=\"1440\" \/>\n\t<meta property=\"og:image:height\" content=\"960\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/webp\" \/>\n<meta name=\"author\" content=\"admin\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"11 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/?p=5657#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/?p=5657\"},\"author\":{\"name\":\"admin\",\"@id\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/#\\\/schema\\\/person\\\/1b8121bfb9a1c55d1fff3fca675fceaa\"},\"headline\":\"The AI industry is getting better at spotting dangerous behavior. It is less clear that labs know how to stop it.\",\"datePublished\":\"2026-08-20T18:13:23+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/?p=5657\"},\"wordCount\":2262,\"commentCount\":0,\"image\":{\"@id\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/?p=5657#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/wp-content\\\/uploads\\\/2026\\\/08\\\/30b83edc58f3c6db7dbb797db5b1c283.webp\",\"articleSection\":[\"Eye on AI\"],\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/crosscountrymovingteams.com\\\/?p=5657#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/?p=5657\",\"url\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/?p=5657\",\"name\":\"The AI industry is getting better at spotting dangerous behavior. It is less clear that labs know how to stop it. - Cross Country Moving Team\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/?p=5657#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/?p=5657#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/wp-content\\\/uploads\\\/2026\\\/08\\\/30b83edc58f3c6db7dbb797db5b1c283.webp\",\"datePublished\":\"2026-08-20T18:13:23+00:00\",\"author\":{\"@id\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/#\\\/schema\\\/person\\\/1b8121bfb9a1c55d1fff3fca675fceaa\"},\"breadcrumb\":{\"@id\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/?p=5657#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/crosscountrymovingteams.com\\\/?p=5657\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/?p=5657#primaryimage\",\"url\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/wp-content\\\/uploads\\\/2026\\\/08\\\/30b83edc58f3c6db7dbb797db5b1c283.webp\",\"contentUrl\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/wp-content\\\/uploads\\\/2026\\\/08\\\/30b83edc58f3c6db7dbb797db5b1c283.webp\",\"width\":1200,\"height\":600},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/?p=5657#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"The AI industry is getting better at spotting dangerous behavior. It is less clear that labs know how to stop it.\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/#website\",\"url\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/\",\"name\":\"Cross Country Moving Team\",\"description\":\"\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/#\\\/schema\\\/person\\\/1b8121bfb9a1c55d1fff3fca675fceaa\",\"name\":\"admin\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/50b1ad2e498f523425ee0a8cc5180a210646db1622662a3d56cc405d3e0c346a?s=96&d=mm&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/50b1ad2e498f523425ee0a8cc5180a210646db1622662a3d56cc405d3e0c346a?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/50b1ad2e498f523425ee0a8cc5180a210646db1622662a3d56cc405d3e0c346a?s=96&d=mm&r=g\",\"caption\":\"admin\"},\"sameAs\":[\"http:\\\/\\\/crosscountrymovingteams.com\"],\"url\":\"https:\\\/\\\/crosscountrymovingteams.com\\\/?author=1\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"The AI industry is getting better at spotting dangerous behavior. It is less clear that labs know how to stop it. - Cross Country Moving Team","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/crosscountrymovingteams.com\/?p=5657","og_locale":"en_US","og_type":"article","og_title":"The AI industry is getting better at spotting dangerous behavior. It is less clear that labs know how to stop it. - Cross Country Moving Team","og_description":"A new assessment finds leading labs are better at spotting risky behavior than reliably stopping it.","og_url":"https:\/\/crosscountrymovingteams.com\/?p=5657","og_site_name":"Cross Country Moving Team","article_published_time":"2026-08-20T18:13:23+00:00","og_image":[{"width":1440,"height":960,"url":"https:\/\/crosscountrymovingteams.com\/wp-content\/uploads\/2026\/08\/15a8d0a34bb674a57493b6216dfdde1c.webp","type":"image\/webp"}],"author":"admin","twitter_card":"summary_large_image","twitter_misc":{"Written by":"admin","Est. reading time":"11 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/crosscountrymovingteams.com\/?p=5657#article","isPartOf":{"@id":"https:\/\/crosscountrymovingteams.com\/?p=5657"},"author":{"name":"admin","@id":"https:\/\/crosscountrymovingteams.com\/#\/schema\/person\/1b8121bfb9a1c55d1fff3fca675fceaa"},"headline":"The AI industry is getting better at spotting dangerous behavior. It is less clear that labs know how to stop it.","datePublished":"2026-08-20T18:13:23+00:00","mainEntityOfPage":{"@id":"https:\/\/crosscountrymovingteams.com\/?p=5657"},"wordCount":2262,"commentCount":0,"image":{"@id":"https:\/\/crosscountrymovingteams.com\/?p=5657#primaryimage"},"thumbnailUrl":"https:\/\/crosscountrymovingteams.com\/wp-content\/uploads\/2026\/08\/30b83edc58f3c6db7dbb797db5b1c283.webp","articleSection":["Eye on AI"],"inLanguage":"en-US","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/crosscountrymovingteams.com\/?p=5657#respond"]}]},{"@type":"WebPage","@id":"https:\/\/crosscountrymovingteams.com\/?p=5657","url":"https:\/\/crosscountrymovingteams.com\/?p=5657","name":"The AI industry is getting better at spotting dangerous behavior. It is less clear that labs know how to stop it. - Cross Country Moving Team","isPartOf":{"@id":"https:\/\/crosscountrymovingteams.com\/#website"},"primaryImageOfPage":{"@id":"https:\/\/crosscountrymovingteams.com\/?p=5657#primaryimage"},"image":{"@id":"https:\/\/crosscountrymovingteams.com\/?p=5657#primaryimage"},"thumbnailUrl":"https:\/\/crosscountrymovingteams.com\/wp-content\/uploads\/2026\/08\/30b83edc58f3c6db7dbb797db5b1c283.webp","datePublished":"2026-08-20T18:13:23+00:00","author":{"@id":"https:\/\/crosscountrymovingteams.com\/#\/schema\/person\/1b8121bfb9a1c55d1fff3fca675fceaa"},"breadcrumb":{"@id":"https:\/\/crosscountrymovingteams.com\/?p=5657#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/crosscountrymovingteams.com\/?p=5657"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/crosscountrymovingteams.com\/?p=5657#primaryimage","url":"https:\/\/crosscountrymovingteams.com\/wp-content\/uploads\/2026\/08\/30b83edc58f3c6db7dbb797db5b1c283.webp","contentUrl":"https:\/\/crosscountrymovingteams.com\/wp-content\/uploads\/2026\/08\/30b83edc58f3c6db7dbb797db5b1c283.webp","width":1200,"height":600},{"@type":"BreadcrumbList","@id":"https:\/\/crosscountrymovingteams.com\/?p=5657#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/crosscountrymovingteams.com\/"},{"@type":"ListItem","position":2,"name":"The AI industry is getting better at spotting dangerous behavior. It is less clear that labs know how to stop it."}]},{"@type":"WebSite","@id":"https:\/\/crosscountrymovingteams.com\/#website","url":"https:\/\/crosscountrymovingteams.com\/","name":"Cross Country Moving Team","description":"","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/crosscountrymovingteams.com\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Person","@id":"https:\/\/crosscountrymovingteams.com\/#\/schema\/person\/1b8121bfb9a1c55d1fff3fca675fceaa","name":"admin","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/50b1ad2e498f523425ee0a8cc5180a210646db1622662a3d56cc405d3e0c346a?s=96&d=mm&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/50b1ad2e498f523425ee0a8cc5180a210646db1622662a3d56cc405d3e0c346a?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/50b1ad2e498f523425ee0a8cc5180a210646db1622662a3d56cc405d3e0c346a?s=96&d=mm&r=g","caption":"admin"},"sameAs":["http:\/\/crosscountrymovingteams.com"],"url":"https:\/\/crosscountrymovingteams.com\/?author=1"}]}},"_links":{"self":[{"href":"https:\/\/crosscountrymovingteams.com\/index.php?rest_route=\/wp\/v2\/posts\/5657","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/crosscountrymovingteams.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/crosscountrymovingteams.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/crosscountrymovingteams.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/crosscountrymovingteams.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=5657"}],"version-history":[{"count":0,"href":"https:\/\/crosscountrymovingteams.com\/index.php?rest_route=\/wp\/v2\/posts\/5657\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/crosscountrymovingteams.com\/index.php?rest_route=\/wp\/v2\/media\/5655"}],"wp:attachment":[{"href":"https:\/\/crosscountrymovingteams.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=5657"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/crosscountrymovingteams.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=5657"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/crosscountrymovingteams.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=5657"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}