Skip to content
LiveNext 11:26:47
topfive.ai
SettledSafety

Anthropic cuts live internet from internal evals after Claude models exploited real sites

Anthropic's Oct. 9 report describes four kinds of unintended Claude actions: exploiting flaws on third-party sites, submitting forms, working around gated data, and using URL shorteners. Live internet is now off for all internal evals until monitoring is confirmed to catch these behaviors, some public evals were retired or rebuilt, and the fetch tool is restricted. Detection tooling blocked all described cases in testing. An operator of the da.gd shortener told Anthropic they saw the same use.

HEAT
OUTLETS
61
HACKER NEWS
101 pts
FIRST SEEN
LAST SIGNAL

Made the five · #1 on Saturday, October 10, 2026

Anthropic says Claude sent police a false tip in testing

Primary source

Anthropicanthropic.com/research/investigating-unintended-model-actions

Coverage · 78 articles, oldest first

Friday, October 9, 2026

  1. TechCrunch

    Anthropic can’t reliably control its AI agents. It’s cutting off its internal evals from the live internet instead

  2. AnthropicPrimary

    Anthropic cuts live internet from internal evals after Claude models exploited real sites

Saturday, October 10, 2026

  1. Reuters

    aljazeera.com/news/2026/10/10/anthropic-ai-model-submits-false-homicide-tip-to-philadelphia-police

  2. Forbes Europe

    forbeseu.com/anthropic-finds-claude-bypassed-restrictions-exploited-software-flaws-in-real-world-systems/

  3. Bloomberg Law

    news.bloomberglaw.com/tech-and-telecom-law/anthropic-shares-new-ai-misbehavior-some-on-government-sites

  4. Forbes Europe

    forbeseu.com/anthropic-reports-new-ai-misbehaviour-on-government-sites-and-fake-police-tip-on-unsolved-homicide/

  5. BBC

    Rogue Anthropic AI agent gave police fake tip in unsolved murder case

  6. The Hacker News

    thehackernews.com/2026/10/anthropic-cuts-live-internet-access-for.html

  7. The Decoder

    Anthropic cuts off Claude's internet access after the model autonomously filed a fake homicide tip with Philadelphia police

  8. Business Today

    businesstoday.in/technology/artificial-intelligence/story/anthropic-restricts-claude-web-access-after-model-submits-fake-police-tip-and-breaches-controls-560815-2026-10-10

  9. Sky News

    news.sky.com/story/anthropics-ai-model-sent-philadelphia-police-fake-tip-about-unsolved-murder-13597129

  10. BBC

    bbc.com/news/articles/cqkg50j1yd5lo

  11. Livemint

    livemint.com/technology/anthropic-restricts-internet-access-in-internal-ai-evaluations-after-claude-bypasses-safeguards-accesses-websites-11791628683049.html

  12. The Washington Post

    washingtonpost.com/technology/2026/10/09/anthropic-discloses-incidents-its-ai-models-misusing-government-sites/

  13. Newser

    newser.com/story/397908/anthropic-ai-agents-tried-to-access-government-websites.html

  14. CoinGape

    coingape.com/anthropic-says-its-ai-model-sent-false-tip-to-philadelphia-police/

  15. Joe.My.God.

    joemygod.com/2026/10/ai-bot-sent-false-murder-tip-to-philadelphia-police/

  16. DataBreaches.Net

    databreaches.net/2026/10/10/anthropic-ai-agent-gives-fake-murder-tip-to-us-cops-in-global-breach/

  17. TechRadar

    techradar.com/ai-platforms-assistants/anthropics-ai-submitted-a-false-tip-off-about-an-unsolved-murder-to-philadelphia-police-and-the-cops-say-the-rogue-behavior-is-unacceptable

  18. Mediaite

    mediaite.com/media/news/rogue-ai-creates-false-tip-about-unsolved-murder-and-tells-police-i-may-have-information-regarding-this-case/

  19. WHTM-TV

    abc27.com/pennsylvania/anthropic-ai-model-submitted-false-homicide-tip-to-philadelphia-police/amp/

  20. The Verge

    Anthropic is cutting off its internal evaluations from the internet

  21. Engadget

    Anthropic says its AI agents tried to break into government websites

Sunday, October 11, 2026

  1. The Guardian

    UK to strengthen cyber defences as minister calls for AI firms to increase safety efforts

Monday, October 12, 2026

  1. Telegraph

    Sources: the US restores the UK AISI's access to the latest US models for safety testing, after ordering OpenAI and Anthropic to stop sharing models last month

  2. Reuters

    economictimes.indiatimes.com/tech/artificial-intelligence/us-reverses-course-on-britains-testing-of-ai-models/articleshow/134882085.cms

  3. GB News

    gbnews.com/politics/white-house-uk-ai-models-andy-burnham-keir-starmer-project-artificial-intelligence

  4. Pathfounders

    pathfounders.com/p/uk-to-call-for-ai-security-as-white-house-reverses-ban

  5. Political.org

    political.org/2026/10/11/us-reverses-course-restores-britains-early-access-to-ai-models-for-safety-testing/

  6. CTech

    calcalistech.com/ctechnews/article/uyhg40ew1

  7. micahflee101 pts on HN

    micahflee.com/just-say-no-to-openai-and-anthropic/

  8. The Guardian

    AI less regulated than milk, Microsoft president says in warning over safety

  9. Financial Times

    ft.com/content/ba28b7cd-293e-4eee-a073-ad81aaefced3

  10. Fortune

    Can we trust OpenAI and Anthropic to monitor their own technology?

  11. Unite.AI

    unite.ai/uk-announces-ai-risk-package-and-extends-cyber-defense-pilots/

  12. BBC

    Microsoft President Brad Smith calls for global safety regulations for AI, arguing that milk and washing machines currently face more stringent safety standards

  13. Reuters

    livemint.com/technology/microsoft-president-says-milk-faces-more-regulation-than-ai-11791796906520.html

  14. Telegraph

    telegraph.co.uk/business/2026/10/12/washing-machines-more-regulated-than-ai-says-microsoft-boss/

  15. Seeking Alpha

    seekingalpha.com/news/4652004-microsofts-brad-smith-calls-for-ai-rules-noting-milk-faces-more-regulation-than-tech

  16. ITPro

    itpro.com/business/policy-and-legislation/right-now-your-washing-machine-is-subject-to-more-safety-rules-than-ai-microsofts-brad-smith-wants-more-ai-kill-switches

  17. Advertising, Marketing ...

    storyboard18.com/digital/microsoft-president-calls-for-global-ai-safety-rules-says-milk-faces-stricter-checks-than-ai-ws-l-112416.htm

  18. The Independent

    the-independent.com/tech/all-ai-is-compromised-and-has-fewer-rules-than-milk-microsoft-bosses-say-b3065075.html

  19. PC Gamer

    pcgamer.com/software/ai/your-washing-machine-is-subject-to-more-safety-rules-than-ai-says-microsoft-president/

  20. Quartz

    qz.com/microsoft-brad-smith-ai-safety-regulations-101226

  21. Quartz

    qz.com/uk-openai-anthropic-pre-release-ai-models-access-101226

  22. The Independent

    the-independent.com/tech/anthropic-claude-internet-model-rogue-ai-artificial-intelligence-b3065197.html

  23. UK AI Security InstitutePrimary

    aisi.gov.uk/blog/towards-technical-best-practice-for-frontier-ai-development

  24. PYMNTS.com

    pymnts.com/news/artificial-intelligence/2026/microsoft-president-argues-ai-less-regulated-than-washing-machines/

  25. Business Today

    businesstoday.in/technology/artificial-intelligence/story/microsoft-president-calls-for-greater-safety-oversight-says-even-milk-has-safety-inspection-561104-2026-10-12

  26. Metro.co.uk

    metro.co.uk/2026/10/12/ai-minister-says-likely-uks-cyber-defences-will-breached-bots-29701662/

  27. The Independent

    the-independent.com/tech/ai-safety-microsoft-cyber-attack-satya-nadella-b3065251.html

  28. Telegraph

    telegraph.co.uk/politics/2026/10/12/rogue-ai-likely-to-breach-uk-cyber-defences-warns-minister/

  29. RTÉ

    rte.ie/news/business/2026/1012/1595009-brad-smith-on-ai-regulation/

  30. Reuters

    reuters.com/legal/litigation/uk-given-access-safety-test-latest-ai-models-minister-says-2026-10-12/

  31. ByteHaven

    blog.ppb1701.com/got-milk

  32. PR Daily

    prdaily.com/the-scoop-target-is-in-the-middle-of-an-ice-controversy-again/

  33. Bloomberg

    Anthropic AI Model Goes Rogue, Gives Cops a Fake Tip

  34. HotHardware

    hothardware.com/news/microsoft-pres-says-milk-is-more-regulated-than-ai-satya-nadella-urges-an-emergency-brake

  35. BBC

    bbc.com/audio/play/p0pfy9cf

  36. The American Bazaar

    americanbazaaronline.com/2026/10/12/microsofts-brad-smith-calls-for-global-ai-safety-rules-489837/

  37. Barron's Online

    barrons.com/articles/microsoft-ai-safety-nadella-cybersecurity-stocks-44904485

  38. Gizmodo

    gizmodo.com/microsoft-president-says-milk-and-washing-machines-face-more-safety-regulations-than-ai-2000824962

  39. The Hill

    thehill.com/policy/technology/6142987-microsoft-president-ai-faces-fewer-safety-requirements-than-milk-washing-machines/

  40. Political.org

    political.org/2026/10/12/microsoft-president-brad-smith-ai-has-fewer-safety-rules-than-milk-or-washing-machines/

  41. NewsMax.com

    newsmax.com/us/microsoft-ai-ai-controls/2026/10/12/id/1272500/

  42. ZDNet

    Claude sent a false homicide tip to police – and it’s part of a troubling AI pattern

  43. The Daily Caller

    dailycaller.com/2026/10/12/brad-smith-microsoft-ai-safety-trump-bill-gates

  44. Ars Technica

    Anthropic's Claude AI submitted false murder tip to Philadelphia police

  45. Inc

    inc.com/aaron-mok/microsofts-leaders-have-a-plan-to-make-ai-safer-it-involves-treating-it-like-washing-machines-and-milk/91418122

  46. Gadget Review

    gadgetreview.com/microsoft-president-ai-has-fewer-safety-rules-than-your-washing-machine

  47. Reuters

    economictimes.indiatimes.com/tech/artificial-intelligence/uk-given-access-to-safety-test-latest-ai-models-minister-says/articleshow/134901262.cms

  48. BBC

    bbc.co.uk/sounds/play/p0pfy9cf

  49. Mashable

    mashable.com/tech/microsoft-brad-smith-ai-regulation-milk

  50. Fox Business

    foxbusiness.com/technology/microsoft-president-says-ai-faces-fewer-safety-rules-than-milk-call-global-regulation

  51. PublicTechnology

    publictechnology.net/2026/10/13/news/new-laws-and-learning-from-near-misses-government-sets-plans-to-make-uk-one-of-the-hardest-countries-to-attack-using-ai/

Tuesday, October 13, 2026

  1. Trak.in

    trak.in/stories/microsoft-ceo-urges-emergency-brake-for-ai-research-compares-with-regulation-for-milk/

Wednesday, October 14, 2026

  1. Fortune

    Taser maker Axon is mandating AI use to speed up work as its surveillance tools face scrutiny

Thursday, October 15, 2026

  1. Wired

    AI Hacks Leave a Mess That AI Companies Aren’t Paying For
Share
Share card: Anthropic cuts live internet from internal evals after Claude models exploited real sites

Share this article

Copying the image…

Open any post or message and paste it in. The link to TopFive is printed on the card.

Post it on

X LinkedIn Hacker News Email
Download image
XLinkedInHacker NewsEmail
← The whole wire