Window ending Oct. 12, 2026

This is a backcast from the current catalogue. It was not a reading published at the time.

Records behind this reading

21 selected records; 6 documented external harms counting (7 qualifying). Records and ratings below reflect the current catalogue.

6 harms counting that week: 1 minor, 5 negligible (3 weighted for control). 21 records reported in the last 30 days.

Records behind this reading · 6

• PI-0090 · AI coding agents posted 13,000 internal screenshots from over 300 companies to public GitHub repositoriesMinor harm, reported 1 week ago · bounded single-source review · extent undisclosed
• PI-0092 · Claude Mythos Preview exploited an injection flaw to run commands on a university server during an evaluationNegligible harm, reported 3 days ago, counts as 2 records: it reached outside systems · extent undisclosed
• PI-0081 · OpenAI agents edited Wikimedia wikis without approval and tried to use a citation tool as a proxy, the foundation saysNegligible harm, reported 1 week ago
• PI-0083 · An OpenAI agent reached non-public statistics in a New South Wales parks agency's fire history serviceNegligible harm, reported 1 week ago, counts as 2 records: it reached outside systems · extent undisclosed
• PI-0067 · OpenAI research agents posted user-provided images to image-hosting sites in 53 instancesMinor harm, reported 2 weeks ago, now counting as negligible · extent undisclosed
• PI-0064 · OpenAI research agent got around controls on an Australian government Medicare statistics portalMinor harm, reported 2 weeks ago, now counting as negligible, counts as 2 records: it reached outside systems · extent undisclosed

1 internal harm excluded.

Only these records move the reading. How it is calculated

1. PI-0061
   OpenAI agents put task files on the public internet against instructionsSept. 16, 2026 · No harm reported · Internal research
2. PI-0062
   OpenAI model used a leaked third-party API key, then fabricated the data it could not fetchSept. 16, 2026 · No harm reported · Internal research
3. PI-0063
   OpenAI models wrote notes telling future copies to hide mistakes and ignore constraintsSept. 16, 2026 · No harm reported · Internal research
4. PI-0078
   OpenAI training agents used an internal package repository as a message board across separate samplesSept. 16, 2026 · No harm reported · Internal research
5. PI-0073
   Gemini guessed a password and used leaked credentials to get into three real companies during a hacking testSept. 18, 2026 · Negligible harm · no longer counting · Controlled test
6. PI-0077
   Early Claude Opus 5.5 snapshot wrote a command to send secrets to an external host during internal useSept. 22, 2026 · No harm reported · Internal research
7. PI-0065
   Researchers linked attempted break-ins at Data USA and a university library to OpenAI agentsSept. 23, 2026 · No harm found · Internal research
8. PI-0070
   AI research agents scanned a UN statistics API about 16,500 times and worked around its request limitsSept. 23, 2026 · No harm reported · Internal research
9. PI-0064
   OpenAI research agent got around controls on an Australian government Medicare statistics portalSept. 24, 2026 · Minor harm · counting as negligible, rated minor · Internal research
10. PI-0075
    In a researcher demo, a planted web lead hijacked Salesforce Agentforce into leaking account data through DNSSept. 24, 2026 · No harm reported · Controlled test
11. PI-0066
    OpenAI internal model leaked a researcher's GitHub token into a public repository while trying to cheatSept. 25, 2026 · Negligible harm · Internal research
12. PI-0067
    OpenAI research agents posted user-provided images to image-hosting sites in 53 instancesSept. 25, 2026 · Minor harm · counting as negligible, rated minor · Internal research
13. PI-0068
    OpenAI training agent bypassed network controls to reach an outside chatbotSept. 25, 2026 · No harm reported · Internal research
14. PI-0069
    OpenAI agents pulled data from US government sites; researchers say one tried to hack an Education Department siteSept. 25, 2026 · No harm found · Internal research
15. PI-0090
    AI coding agents posted 13,000 internal screenshots from over 300 companies to public GitHub repositoriesSept. 29, 2026 · Minor harm · Deployment
16. PI-0082
    In a researcher demo, a hidden email made the Manus agent run attacker code past its prompt-injection guardOct. 1, 2026 · No harm reported · Controlled test
17. PI-0086
    GPT-6 Astra downloaded a top human-written bot and tried to enter it as its own in the StarSkirmish StarCraft benchmarkOct. 2, 2026 · No harm reported · Controlled test
18. PI-0083
    An OpenAI agent reached non-public statistics in a New South Wales parks agency's fire history serviceOct. 3, 2026 · Negligible harm · Internal research
19. PI-0081
    OpenAI agents edited Wikimedia wikis without approval and tried to use a citation tool as a proxy, the foundation saysOct. 5, 2026 · Negligible harm · Internal research
20. PI-0091
    Claude Opus 5 and Mythos 5 used free URL shorteners to get around their fetch tool's URL length limitOct. 9, 2026 · No harm reported · Internal research
21. PI-0092
    Claude Mythos Preview exploited an injection flaw to run commands on a university server during an evaluationOct. 9, 2026 · Negligible harm · Controlled test

Sources

These links support the records above. Source availability and conclusions may change.

1. OpenAI: Our framework for reporting model misalignmentCited in PI-0061, PI-0062, PI-0063
2. OpenAI Alignment: Uploading files to the internet in order to cite themCited in PI-0061
3. OpenAI Alignment: Unauthorized communication via temporary file hosting servicesCited in PI-0061
4. The Hacker News: OpenAI Reveals Six Model Incidents Involving Hidden Failures and Unauthorized UploadsCited in PI-0061
5. NBC News: OpenAI flags 6 new incidents of 'concerning' behavior and unveils plan to track itCited in PI-0061
6. OpenAI Alignment: Signing up for disposable emails and searching GitHub for leaked API keysCited in PI-0062
7. SecurityWeek: OpenAI Says Its Models Searched GitHub for Leaked API Keys During TrainingCited in PI-0062
8. Dark Reading: Rogue Behavior: OpenAI Reveals More Model Misalignment IncidentsCited in PI-0062
9. OpenAI Alignment: Self-generated prompt injections in compaction summariesCited in PI-0063
10. OpenAI Alignment: Encouraging deception in compaction summariesCited in PI-0063
11. Axios: OpenAI discloses six new AI misalignment incidentsCited in PI-0063
12. BleepingComputer: OpenAI details more cases of AI agents taking unauthorized actionsCited in PI-0063
13. OpenAI Alignment: Unsanctioned Artifactory writes and cross-sample communicationCited in PI-0078
14. CSO Online: OpenAI admits six new misalignment incidents under new reporting frameworkCited in PI-0078
15. Irregular: Addressing Recent Incidents: Ongoing Findings and Path ForwardCited in PI-0073
16. TechRadar: Google's Gemini hacked three companies during Irregular AI 'capture-the-flag' testingCited in PI-0073
17. Cybernews: Google's Gemini hacked three real companies during security testCited in PI-0073
18. Al Jazeera (Reuters): Google's Gemini AI hacks 3 companies in security test, then stopsCited in PI-0073
19. GV Wire (Reuters): Gemini Hacked Three Companies in First Known Breakout By Google's AI, WSJ reportsCited in PI-0073
20. Anthropic: System Card: Claude Opus 5.5Cited in PI-0077
21. MIXED: Pre-release Opus 5.5 wrote secret-stealing commands after a copying slip, says AnthropicCited in PI-0077
22. Transluce: Early rogue AI agent activity and attempts to hack found on urlquery.netCited in PI-0065, PI-0070
23. TechCrunch: For months, OpenAI's agent swarms have been attacking online databases to find obscure factsCited in PI-0065
24. infoDOCKET (Library Journal): OpenAI Agent Hacking Attempt Targets University of New Mexico Digital LibraryCited in PI-0065
25. KOB 4: OpenAI agent hacking attempt targets University of New Mexico digital libraryCited in PI-0065
26. Rowan Howard-Jones (swarmcha.se): OpenAI agents tried to bruteforce a UN website's API fieldsCited in PI-0070
27. The Register: OpenAI agents went the long way round for UN dataCited in PI-0070
28. Quartz: OpenAI's AI agents hit a UN website 16,000 times — bypassing its security filtersCited in PI-0070
29. The Next Web: OpenAI agents scanned a UN statistics site 16,500 times, researcher saysCited in PI-0070
30. OpenAI: How we will do better for AustraliaCited in PI-0064
31. Prime Minister of Australia: Press conference – New YorkCited in PI-0064
32. ABC News (Australia): OpenAI agent hacked Medicare portal, PM saysCited in PI-0064
33. CNBC: OpenAI says agent hacked Australian government website without being told to do soCited in PI-0064
34. TIME: Australia Condemns 'Unacceptable' OpenAI Breach of Government Health PortalCited in PI-0064
35. BleepingComputer: OpenAI hacked Australian Medicare govt site, probed data providersCited in PI-0064
36. Zenity Labs: SalesBleed: Indirect Prompt Injection and 0-Click Data Exfiltration on AgentforceCited in PI-0075
37. OpenAI Alignment: Exposing a GitHub token in a public repositoryCited in PI-0066
38. TechCrunch: OpenAI still doesn't seem to have a handle on all of its rogue AI activityCited in PI-0066
39. OpenAI: The Hugging Face incident and other third-party impact from misaligned models (September 25 update)Cited in PI-0067
40. TechCrunch: Unsecured OpenAI agents posted 53 user images on the internet without the lab's knowledgeCited in PI-0067
41. BleepingComputer: OpenAI's AI agents accidentally uploaded user-provided images to third-party sitesCited in PI-0067
42. Newsweek: OpenAI Admits AI Agents Exposed 53 User Images During ResearchCited in PI-0067
43. OpenAI Alignment: An agent used DNS to reach an external chatbotCited in PI-0068
44. The Hacker News: OpenAI Pauses Tool Use After Agent Bypasses Internet Controls to Reach External ChatbotCited in PI-0068
45. Fortune: OpenAI pauses training a second time after saying its AI agents escaped a secure 'sandbox' againCited in PI-0068
46. The Register: OpenAI pauses some training amid allegations its rogue agents behaved more badly than first thoughtCited in PI-0068
47. Transluce: AI Agents Targeted U.S. and Canadian Government WebsitesCited in PI-0069
48. NPR: OpenAI says its models engaged with US government websites in misbehavior disclosureCited in PI-0069
49. CNN Business: Rogue OpenAI agents targeted three separate US government websitesCited in PI-0069
50. The Hill: OpenAI agent unsuccessfully tried to breach Department of Education websiteCited in PI-0069
51. Nextgov/FCW: OpenAI agents accessed Census, SEC data and tried to hack Education websiteCited in PI-0069
52. Education Week (AP): OpenAI's Models Probed Websites of Department of Education, Other AgenciesCited in PI-0069
53. Glow Security: PixelLeak: How AI Agents Exposed Developer Screenshots from Leading Tech CompaniesCited in PI-0090
54. The Register: AI models keep posting screenshots showing sensitive data from inside tech companiesCited in PI-0090
55. The Hacker News: AI Coding Agents Exposed 13,000 Internal Images, Including Billing Records, on GitHubCited in PI-0090
56. Salt Labs: How We Hijacked an AI Agent With a Single EmailCited in PI-0082
57. TechRadar: This popular AI agent could be hacked by a single email — with potentially disastrous consequencesCited in PI-0082
58. heise online: StarCraft benchmark: GPT-6 Astra cheats with a foreign botCited in PI-0086
59. PC Gamer: An OpenAI model was caught trying to cheat at StarCraft, and of course it did it by stealing a human's workCited in PI-0086
60. Destructoid: When an AI couldn’t win in StarCraft, it did the next-best thing: downloaded a human-made botCited in PI-0086
61. 7NEWS: OpenAI slammed after another government agency hacked by AI agentCited in PI-0083
62. Wikimedia Foundation: OpenAI "rogue" agent activities found on Wikimedia projectsCited in PI-0081
63. BleepingComputer: Wikimedia: Rogue OpenAI agents behind unauthorized Wikipedia editsCited in PI-0081
64. The Hacker News: Wikimedia Says OpenAI Agents Tried to Compromise Etherpad and Use Wiki Tools as ProxiesCited in PI-0081
65. The Next Web: 'The open web is a public good': Wikimedia on rogue OpenAI agentsCited in PI-0081
66. Reuters (via Dawn): Wikipedia operator says OpenAI's rogue agents possibly tied to data service disruption in MayCited in PI-0081
67. Anthropic: Investigating unintended model actions in our evaluations and internal useCited in PI-0091, PI-0092

Every weekly brief · How this reading is calculated · Download the weekly card