{"slug": "ai-labs-are-failing-to-keep-their-own-systems-in-check", "title": "AI labs are failing to keep their own systems in check", "summary": "No AI company fully applies basic control measures to its own internal AI systems, according to the first assessment by nonprofit Guidelight, which graded Anthropic and OpenAI with a C+, Google with a D+, xAI with a D−, and Meta with an F. The group evaluated six practices including logging, review mechanisms, circuit breaking, and containment plans, finding that companies do best at spotting misbehavior but worst at prevention and containment.", "body_md": "# AI labs are failing to keep their own systems in check\n\n**No AI company fully applies basic control measures to its own internal AI systems.** That's the takeaway from the first assessment by the nonprofit Guidelight. The group looked at Anthropic, OpenAI, Google, xAI, and Meta, drawing only on public sources like system cards, safety reports, and blog posts.\n\nGuidelight checked six basic practices. These include logging internal AI activity, gating risky actions through a review mechanism, emergency shutdowns known as \"circuit breaking,\" and plans to contain misaligned models. Anthropic and OpenAI lead with a C+, Google follows with a D+ and a detailed [roadmap](https://arxiv.org/pdf/2607.13087), while xAI (D−) and Meta (F) score the worst.\n\nThe companies do best at spotting misbehavior. They do worst at prevention and containment. Guidelight is an independent nonprofit founded by former OpenAI safety leads Page Hedley and Steven Adler.\n\n```\nAI News Without the Hype – Curated by Humans\n\n\t\t\t\t\tSubscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive \"AI Radar\" frontier report six times a year, full archive access, and access to our comment section.\t\t\t\t\n\n\t\t\t\t\tSubscribe now\n```\n\n[Guidelight](https://guidelight.ai/blog/control-assessment-august-2026)", "url": "https://wpnews.pro/news/ai-labs-are-failing-to-keep-their-own-systems-in-check", "canonical_source": "https://the-decoder.com/ai-labs-are-failing-to-keep-their-own-systems-in-check/", "published_at": "2026-08-19 13:16:18+00:00", "updated_at": "2026-08-19 13:43:49.481461+00:00", "lang": "en", "topics": ["ai-safety", "ai-policy"], "entities": ["Guidelight", "Anthropic", "OpenAI", "Google", "xAI", "Meta", "Page Hedley", "Steven Adler"], "alternates": {"html": "https://wpnews.pro/news/ai-labs-are-failing-to-keep-their-own-systems-in-check", "markdown": "https://wpnews.pro/news/ai-labs-are-failing-to-keep-their-own-systems-in-check.md", "text": "https://wpnews.pro/news/ai-labs-are-failing-to-keep-their-own-systems-in-check.txt", "jsonld": "https://wpnews.pro/news/ai-labs-are-failing-to-keep-their-own-systems-in-check.jsonld"}}