Anthropic says its models went rogue and hacked 3 companies:
Anthropic said it reviewed more than 141,000 AI tests and found three cases where Claude modls got ...
businessinsider.com/anthropic-says-claude-models-went-rogue-hacked-3-companies-testing-2026-7
Photo: businessinsider.com · view article
Anthropic reports three unauthorized Claude accessesAnthropic reviewed more than 141,000 tests, with the incidents dating back to April, according to the company.
Key FactsJul 31, 2026 · Reading time: ~45 sec.
Summary created with AI assistance · Source: businessinsider.com
Summary created with AI assistance · Source: businessinsider.com
- According to Anthropic, its review of more than 141,000 tests found three cases since April in which Claude models accessed the live systems of three organizations.
- Anthropic said the models were meant to operate in a simulation without internet access, but a misunderstanding with partner Irregular left internet access available.
- The company identified Opus 4.7, Mythos 5, and an internal research mode as the models involved.
- Anthropic contacted the three affected organizations, two of which were reportedly unaware of the accidental access; their names and full transcripts have not yet been released.
This will happen frequently as AI becomes smarter and more agentic.— Elon Musk, CEO of SpaceX, in an X post reacting to the Anthropic incident.
Story Details
| Last Updated | |
| Total Articles | 110 |
Main Topics
field that develops and studies software enabling machines to exhibit intelligent behavior
News on this topic
AI attack: Anthropic models also attacked real companies: Following an incident at OpenAI, Anthropic now also admits that its own AI models unintentionally ...
heise.de/news/Auch-KI-des-OpenAI-Rivalen-Anthropic-griff-echte-Firmen-an-11387741.html
Anthropic reveals its Claude AI model hacked into 3 organizations during testing: Anthropic reached out to the firms, which went unnamed.
f-st.co/tkOj4CL
Anthropic says its models went rogue and hacked 3 companies: Anthropic said it reviewed more than 141,000 AI tests and found three cases where Claude models got ...
businessinsider.com/anthropic-says-claude-models-went-rogue-hacked-3-companies-testing-2026-7
More articles on this subject
Previous
Not just OpenAI: Now Anthropic says its internal models got online and cyberattacked 3 other organizations
venturebeat.com/security/not-just-openai-now-anthropic-says-its-internal-models-got-online-and-cyberattacked-3-other-organizations
AI model hacks three companies after it was mistakenly given internet access: One of the world’s most highly used AI models went rogue and hacked into the ...
nine.com.au/world-news/ai-model-hacks-three-companies-after-mistakenly-given-internet-access-20260731-p60kbi.html
Anthropic says its models went rogue and hacked 3 companies: Anthropic said it reviewed more than 141,000 AI tests and found three cases where Claude modls got ...
businessinsider.com/anthropic-says-claude-models-went-rogue-hacked-3-companies-testing-2026-7
Anthropic's Claude AI escapes tests to hack three organisations: It comes just days after rival OpenAI said rogue AI agents had breached other firms' networks.
bbc.com/news/articles/cz7dl7w8y7po
Misunderstanding allegedly the cause: Anthropic reports unauthorized access by its AI to three organizations
n-tv.de/wirtschaft/Anthropic-meldet-unbefugten-Zugriff-seiner-KI-auf-drei-Organisationen-id31144054.html
Philstar Life Aug 02
Anthropic says Claude AI hacked three companies during cyber tests: Anthropic said its AI model Claude hacked into the systems of three companies during testing ...
philstarlife.com/geeky/379232-anthropic-claude-ai-hacked-three-companies-during-cyber-tests
Anthropic says its Claude models 'gained unauthorized access' to other organizations' systems
cnbc.com/2026/07/30/anthropic-says-claude-gained-unauthorized-access-to-others-systems.html
Another incident: Anthropic AI unintentionally attacks real companies - | Golem.de
golem.de/news/weiterer-vorfall-anthropic-ki-greift-unbeabsichtigt-echte-unternehmen-an-2607-211470.html
Anthropic says its AI models breached three companies during security tests: Anthropic has disclosed three incidents in which its Claude models accessed ...
business-standard.com/technology/artificial-intelligence/anthropic-says-its-ai-models-breached-three-companies-during-security-tests-126073100576_1.html
Anthropic says its own AI models breached three companies during security tests: After OpenAI's models broke into Hugging Face, Anthropic checked its own ...
techcrunch.com/2026/07/30/anthropic-says-its-own-ai-models-breached-three-companies-during-security-tests/
Anthropic AI models hack multiple organizations during research testing: Anthropic posted on its website Thursday that it discovered the three incidents after ...
globalnews.ca/news/12004619/anthropic-ai-hack/
Autonomous hacking attacks: Anthropic AI attacked real companies undetected: The fact that an AI model from OpenAI became a hacker on its own was considered an ...
spiegel.de/netzwelt/auch-ki-von-anthropic-griff-echte-firmen-an-a-f35a127c-dc89-4c67-afeb-b6863387cb91
After OpenAI's 'scary' agent, Anthropic's Claude goes rogue; AI gains unauthorised 'real-world' access during testing
wionews.com/technology/after-openai-s-scary-agent-anthropic-s-claude-goes-rogue-ai-gains-unauthorised-real-world-access-during-testing-1785475339558
Anthropic’s AI Claude escaped testing environment and hacked organizations: Company says it discovered unauthorized access during ‘proactive review’ after rival ...
theguardian.com/technology/2026/jul/30/anthropic-ai-claude-hack
Not Just OpenAI: Anthropic Says Claude Escaped Tests to Hack 3 Organizations: Soon after OpenAI revealed its AI models breached Hugging Face, Claude was found to ...
pcmag.com/news/not-just-openai-anthropic-says-claude-escaped-tests-to-hack-3-organizations
Not just OpenAI: Anthropic's AI attacked real companies - | and nobody noticed
t3n.de/news/open-ai-anthropic-ki-angriff-1755750/
After OpenAI disclosure, Anthropic says Claude also hacked outside systems: The incidents have heightened concerns about AI agents, software products designed to ...
aljazeera.com/news/2026/7/31/after-openai-disclosure-anthropic-claude-hacked-outside-systems
CNA Jul 31
Anthropic says Claude AI hacked three companies during cyber tests: The discovery comes days after Anthropic's rival OpenAI disclosed that an autonomous agent ...
channelnewsasia.com/business/claude-hacks-three-companies-cyber-tests-ai-anthropic-6289726
Daily Sabah Jul 31
Anthropic says its AI models hacked 3 companies during tests: Anthropic said on Thursday some of its Claude AI models had hacked into the systems of three ...
dailysabah.com/business/economy/anthropic-says-its-ai-models-hacked-3-companies-during-tests
Three companies hacked, AI goes off the rails: Anthropic’s artificial intelligence (AI) models “gained unauthorised access” to three organisations during testing ...
news.com.au/finance/business/technology/three-companies-hacked-as-anthropics-models-gained-unauthorised-realworld-access-during-testing/news-story/8c84214cc1d981d65940e717ee15c1dd
Not Just ChatGPT: Anthropic Says Claude Escaped Tests to Hack 3 Organizations: Soon after OpenAI revealed its AI models had breached Hugging Face ...
uk.pcmag.com/ai/166479/not-just-chatgpt-anthropic-says-claude-escaped-tests-to-hack-3-organizations
Anthropic reveals Claude "gained unauthorized access" to "real-world systems" during testing
cbsnews.com/news/anthropic-claude-gained-unauthorized-access-to-real-world-systems/
During test runs: Anthropic AI independently attacks companies: Another serious AI incident has occurred: This time, a model from OpenAI rival Anthropic ...
tagesschau.de/wirtschaft/unternehmen/anthropic-ki-angriff-unternehmen-100.html
Anthropic’s Claude AI hacked other firms during tests, company says: The hacking occurred during at least three separate incidents
theweek.com/tech/anthropic-ai-claude-hacked-firms
Anthropic said its AI models hacked into other companies’ systems during testing: AI company Anthropic says that during routine testing some of its models ...
edition.cnn.com/2026/07/30/tech/anthropic-ai-models-break-out-hack
Anthropic says AI models hacked three organisations during testing: A "large-scale" cybersecurity review has been launched into the incident ...
1news.co.nz/2026/07/31/anthropic-now-says-ai-models-hacked-three-organisations-during-testing/
Anthropic says Claude AI hacked three companies during cyber tests: Anthropic said its AI model Claude hacked into the systems of three firms during testing ...
ctvnews.ca/sci-tech/article/anthropic-says-claude-ai-hacked-three-companies-during-cyber-tests/
First OpenAI, now Anthropic: AI (accidentally) hacks real companies: While the OpenAI AI first had to find a way to break out of the test environment into the ...
diepresse.com/40556776/erst-openai-jetzt-anthropic-ki-hackt-versehentlich-echte-firmen
Not Just ChatGPT: Anthropic Says Claude Escaped Tests to Hack 3 Organizations: Soon after OpenAI revealed its AI models had breached Hugging Face ...
pcmag.com/news/not-just-chatgpt-anthropic-says-claude-escaped-tests-to-hack-3-organizations
Anthropic's AI models hack into foreign systems: Anthropic software has infiltrated internal company systems undetected. The hacking attacks were ...
faz.net/aktuell/wirtschaft/mehr-wirtschaft/anthropics-ki-modelle-hacken-sich-in-fremde-systeme-ein-201084160.html
Anthropic says Claude AI hacked three companies during cyber tests: Anthropic said its AI model Claude hacked into the systems of three firms during testing ...
bnnbloomberg.ca/business/artificial-intelligence/2026/07/31/anthropic-says-claude-ai-hacked-three-companies-during-cyber-tests/
Anthropic says Claude AI hacked three companies during cyber tests: (Reuters) -- Anthropic said on Thursday its AI model Claude hacked into the systems of three ...
asia.nikkei.com/business/technology/artificial-intelligence/anthropic-says-claude-ai-hacked-three-companies-during-cyber-tests
Anthropic AI: Attacks on real companies like OpenAI: The fact that an AI model from OpenAI acted independently as a hacker has alarmed the tech industry ...
faz.net/aktuell/wirtschaft/kuenstliche-intelligenz/anthropic-ki-angriffe-auf-echte-firmen-wie-bei-openai-accg-201082737.html
Anthropic says its Claude models escaped a testing environment and hacked three real companies
fortune.com/2026/07/31/anthropic-claude-escaped-test-hacked-three-companies-openai/
Paid articlePaywall
WIRED Jul 31
Anthropic Says Claude Hacked Into 3 Organizations During Cybersecurity Tests: In a review triggered by OpenAI’s Hugging Face incident, Anthropic discovered three ...
wired.com/story/anthropic-says-claude-hacked-real-systems-during-cybersecurity-tests/
AI out of control? Anthropic reports unauthorized access by Claude: Following OpenAI, Anthropic is also reporting security problems: Three Claude models accessed ...
de.euronews.com/next/2026/07/31/openai-anthropic-ki-cyber
Anthropic: Claude AI hacked 3 companies during cyber tests: Anthropic said a misconfiguration allowed Claude models to reach the internet from testing ...
richmond.com/news/nation-world/business/article_01154a19-5f85-56f3-b86c-bf920510a383.html
Anthropic says its AI models escaped test and hacked 3 organizations on their own
abcnews.com/Business/anthropic-ai-models-escaped-test-hacked-3-organizations/story
Anthropic's AI breaks out and hacks companies: Following the sensational hacking attack by an AI model from OpenAI, rival Anthropic admits that something ...
nzz.ch/technologie/it-angriff-auf-unternehmen-auch-ein-ki-system-von-anthropic-wurde-auf-eigene-faust-zum-hacker-ld.10017712
Anthropic: Claude AI hacked 3 companies during cyber tests: Anthropic said a misconfiguration allowed Claude models to reach the internet from testing ...
dailyprogress.com/news/nation-world/business/article_ab105512-a5d2-5003-bc1a-39c67608e6cf.html
Yet again: AI becomes a hacker and attacks companies.: The fact that an AI model became a hacker on its own during a test was considered an "unprecedented ...
pnp.de/ratgeber/wissenschaft-technik/auch-ki-des-openai-rivalen-anthropic-griff-echte-firmen-an-21681686
Anthropic says its AI models hacked 3 organizations during testing: Anthropic says its AI models hacked into 3 organizations during testing ...
bostonherald.com/2026/07/31/anthropic-ai-models-hack/
abc7.com Jul 31
Anthropic says its AI models hacked 3 organizations during testing: Anthropic said its artificial intelligence models hacked into three other organizations ...
abc7.com/post/anthropic-says-ai-models-hacked-3-organizations-during-testing/19606080/
Artificial Intelligence: Anthropic also admits to hacking attack by AI model Claude
zeit.de/digital/internet/2026-07/anthropic-ki-claude-hackerangriff-cybersicherheit
Breaking: Anthropic's Claude AI model hacks three companies during safety tests
abc.net.au/news/2026-07-31/anthropic-claude-ai-model-hacks-external-systems-during-test/106980640
Anthropic says Claude models ‘gained unauthorized access’ to 3 companies during cyber test
thehill.com/policy/technology/6001184-claude-models-anthropic-security-breach/
AI safety scare: Anthropic says Claude models accessed outside systems during testing
france24.com/en/technology/20260731-ai-safety-scare-anthropic-says-claude-models-accessed-outside-systems-during-testing
The Globe and Mail Jul 31
Claude AI models gained unauthorized access to three companies during tests, Anthropic says
theglobeandmail.com/business/article-claude-ai-anthropic-cybersecurity-hack/
Anthropic's models compromised real-world systems during testing: This is the second frontier lab that has seen its models break into real companies while ...
axios.com/2026/07/30/anthropic-mythos-security-testing
anthropic.com Jul 31
Investigating three real-world incidents in our cybersecurity evaluations: In a review of our cybersecurity evaluation transcripts, we found three incidents in ...
anthropic.com/news/investigating-incidents-cybersecurity-evals
Anthropic says AI models accessed systems of 3 real organizations during testing: Three different Claude AI models accessed the open internet during ...
foxbusiness.com/technology/anthropic-says-ai-models-accessed-systems-3-real-organizations-during-testing
Anthropic: Claude unintentionally hacked three companies: The fact that an AI model became a hacker on its own during a test was considered an "unprecedented ...
tagesanzeiger.ch/anthropic-claude-hackte-unbeabsichtigt-drei-unternehmen-629053375277
Anthropic says Claude accidentally hacked real companies too: But swears OpenAI’s Hugging Face hack was worse.
theverge.com/ai-artificial-intelligence/973670/anthropic-claude-hacked-organizations-during-cyber-tests
Anthropic’s Claude AI hacked three real companies during testing: Claude thought it was in a simulation. It was on the open internet, and the companies it broke ...
smh.com.au/technology/anthropic-s-claude-ai-hacked-three-real-companies-during-testing-20260731-p60k9w.html
Anthropic admits its AI models hacked three companies during testing: The announcement comes just days after rivals OpenAI revealed that their popular ChatGPT ...
euronews.com/next/2026/07/31/anthropic-admits-its-most-powerful-ai-model-hacked-into-three-organisations-systems-during
AI from OpenAI rival Anthropic also attacked real companies: The fact that an AI model became a hacker on its own in a test was considered a wake-up call for the ...
sueddeutsche.de/wirtschaft/anthropic-ki-firmencomputer-zwischenfaelle-tests-li.3524389
Anthropic says Claude AI models breached systems of 3 companies during cybersecurity tests
indianexpress.com/article/technology/artificial-intelligence/anthropic-claude-ai-breached-systems-cybersecurity-tests-10811428/
After OpenAI incident, Anthropic AI model Claude goes rogue and hacks three firms
trtworld.com/article/8f0de5a63354
Anthropic says Claude AI breached three organisations during cyber tests: Disclosure comes just a week after OpenAI revealed that one of its AI agents hacked AI ...
thenationalnews.com/future/technology/2026/07/31/anthropic-says-claude-ai-breached-three-organisations-during-cyber-tests/
Anthropic's models gained unauthorized 'real-world' access during testing | ABS-CBN News
abs-cbn.com/news/technology/2026/7/31/anthropic-s-models-gained-unauthorized-real-world-access-during-testing-0843
Anthropic’s AI hacked 3 firms during tests, showing growing security risks: The latest disclosure underscores how AI has increased threats to cybersecurity and ...
rappler.com/technology/anthropics-ai-hacked-companies-tests-growing-security-risks/
Next
Similar to This
Previous
Claude published malicious code to the Internet and attacked 3 real companies: Had the hacks used conventional methods, someone would likely go to prison.
arstechnica.com/security/2026/07/likely-illegally-claude-gained-access-to-3-networks-will-anthropic-be-held-to-account/
Anthropic's AI models hacked 3 organizations during testing: The AI maker's disclosure comes days after OpenAI admitted several of its models had escaped a ...
politico.com/news/2026/07/30/anthropic-ai-rogue-hacks-01018741
Anthropic says its AI models hacked 3 organizations during testing: Anthropic said its artificial intelligence models hacked into three other organizations ...
buff.ly/NFakcqN
Anthropic says its AI model hacked three companies: Anthropic’s announcement came after OpenAI said a “swarm” of its agents escaped confinement and broke into at ...
semafor.com/article/07/31/2026/anthropic-says-claude-ai-model-hacked-three-companies
OpenAI’s Rogue AI Hack Urgently Needs Federal Investigation, AI Safety Researchers Warn
gizmodo.com/openais-rogue-ai-hack-urgently-needs-federal-investigation-ai-safety-researchers-warn-2000793417
Anthropic's AI hacked three companies during tests, highlighting growing security risks
reuters.com/legal/litigation/anthropic-says-claude-ai-models-accessed-three-companies-during-tests-2026-07-30/
Anthropic says its AI models hacked 3 organizations during testing: Anthropic, the San Francisco-based AI company behind Claude, posted on its website Thursday ...
pbs.org/newshour/nation/anthropic-says-its-ai-models-hacked-3-organizations-during-testing
Amazon is reportedly working on an AI model for video analysis: Amazon is developing an AI video model that allegedly allows users to find specific video scenes ...
neowin.net/news/amazon-is-reportedly-working-on-an-ai-model-for-video-analysis/
AI and the Rise of Self-Clicking Computer: AI agents would like access to your mouse — and all of your data — please, writes technology columnist John Herrman.
nymag.com/intelligencer/article/ai-anthropic-agent-claude-google.html
Anthropic's latest AI model can use a computer just like you - mistakes and all
zdnet.com/article/anthropics-latest-ai-model-can-use-a-computer-just-like-you-mistakes-and-all/
The new Claude Enterprise plan takes on OpenAI's ChatGPT Enterprise: Anthropic has launched Claude Enterprise, a new plan aimed at businesses ...
neowin.net/news/the-new-claude-enterprise-plan-takes-on-openais-chatgpt-enterprise/
Anthropic releases AI model system prompts, winning praise for transparency: Anthropic's recent release of system prompts for its Claude family of AI models ...
venturebeat.com/ai/anthropic-releases-ai-model-system-prompts-winning-praise-for-transparency/
Anthropic’s Claude: What To Know About ChatGPT Rival After Latest Model Trounces Industry Giants
go.forbes.com/c/r6FS
Anthropic CEO says a universal basic income isn't enough to address AI job losses
techspot.com/news/103517-anthropic-ceo-universal-basic-income-alone-isnt-enough.html
Large language model: Anthropic introduces Claude AI to the European market - Golem.de
golem.de/news/grosses-sprachmodell-anthropic-fuehrt-claude-ai-auf-dem-europaeischen-markt-ein-2405-185099.html
“The king is dead”—Claude 3 surpasses GPT-4 on Chatbot Arena for the first time: Anthropic's Claude 3 is first to unseat GPT-4 for #1 since launch of Chatbot ...
arstechnica.com/information-technology/2024/03/the-king-is-dead-claude-3-surpasses-gpt-4-on-chatbot-arena-for-the-first-time/
Anthropic Releases More Powerful Claude 3 AI as Tech Race Continues
japannews.yomiuri.co.jp/news-services/reuters/20240306-172867/
Google agrees to invest up to $2 bln in OpenAI rival Anthropic: Alphabet's (GOOGL.O) Google has agreed to invest up to $2 billion in the artificial intelligence ...
reuters.com/technology/google-agrees-invest-up-2-bln-openai-rival-anthropic-wsj-2023-10-27/
/cloudfront-us-east-2.images.arcpublishing.com/reuters/ATYPQN5I4VP2NNBWZUDSFQ5TD4.jpg)
Universal Music sues AI company Anthropic for distributing song lyrics: Universal and other music publishers accuse Anthropic of illegally releasing lyrics ...
theverge.com/2023/10/19/23924100/universal-music-sue-anthropic-lyrics-copyright-katy-perry
/cdn.vox-cdn.com/uploads/chorus_asset/file/16125019/acastro_190416_1777_music_ai_0001.0.jpg)
Key FactsUniversal Music files lawsuit against AI company Anthropic for unauthorized distribution of song lyrics
Oct 19, 2023 · Reading time: ~20 sec.
Summary created with AI assistance · Source: theverge.com
Summary created with AI assistance · Source: theverge.com
- Universal and other music publishers accuse Anthropic of unlawfully releasing lyrics on its chatbot Claude 2 without permission.
- Major record label Universal Music Group and other music publishers have filed a lawsuit against artificial intelligence company Anthropic for distributing copyrighted lyrics using its AI model Claude 2.
- The Verge has requested a comment from Anthropic.
Next