Ziel dieser ersten Ausgabe ist eine möglichst breite Bestandsaufnahme der seit dem 4. September auffindbaren Veröffentlichungen zum DseWiki (möglicherweise OpenAI) 'wiki incident'.
BESONDERS RELEVANTE / PRIMÄRE QUELLEN
collusion.wiki / Nightingale Collective u. a. — 4.9.2026
Discovery of a new OpenAI agent message board
https://collusion.wiki/
Primärquelle der Forschergruppe mit Rekonstruktion/Daten.
Reuters — 4.9.2026
EXCLUSIVE: OpenAI agents hijacked German website in previously undisclosed AI breakout this spring
https://www.reuters.com/world/europe/openai-agents-hijacked-german-website-previously-undisclosed-ai-breakout-this-2026-09-04/
Erstbericht mit breiter internationaler Wirkung.
OpenAI auf X — 5.9.2026
How we think about the “wiki incident,” where our agents wrote to several internet sites …
https://x.com/OpenAI/status/2096133504417616165
Offizielle Stellungnahme; OpenAI verwendet selbst die Bezeichnung „wiki incident“ und kündigt Standards für die Offenlegung von Misalignment-Incidents an.
Reuters — 5.9.2026
OpenAI acknowledges ‘wiki incident’ and need for more transparency around unintended AI behavior
https://www.reuters.com/business/media-telecom/openai-acknowledges-wiki-incident-need-more-transparency-around-unintended-ai-2026-09-05/
Folgebericht nach der OpenAI-Stellungnahme.
Reuters — 7.9.2026
OpenAI has sent EU incident report on hijacked German website, Commission says
https://www.reuters.com/business/openai-has-sent-eu-incident-report-hijacked-german-website-commission-says-2026-09-07/
NEUE ENTWICKLUNG: OpenAI hat der EU-Kommission einen Incident Report übermittelt; die Kommission bestätigt laufenden Kontakt.
Euractiv — 7.9.2026
What it feels like when OpenAI’s agents hijack your website
https://www.euractiv.com/news/what-it-feels-like-when-openais-agents-hijack-your-website/
INTERVIEW / BETROFFENENPERSPEKTIVE: ausführliche Darstellung mit Aussagen von Helmut Leitner.
Euractiv — 7.9.2026
AI safety incident reporting ‘not just a tick box’, warns EU
https://www.euractiv.com/news/ai-safety-incident-reporting-not-just-a-tick-box-warns-eu/
REGULATORISCHE REAKTION: EU-Kommission fordert präzise und substanzielle Incident-Berichte.
GROSSE / HOCHWERTIGE MEDIEN
TechCrunch — 4.9.2026
Another swarm of OpenAI agents reached the open internet without the frontier lab’s knowledge
https://techcrunch.com/2026/09/04/another-swarm-of-openai-agents-reached-the-open-internet-without-the-frontier-labs-knowledge/
TechCrunch — 4.9.2026
OpenAI’s rogue agents keep escaping, with no formal process to investigate them
https://techcrunch.com/2026/09/04/openais-rogue-agents-keep-escaping-with-no-formal-process-to-investigate-them/
Eigene Analyse mit Fokus auf unabhängige Incident-Untersuchungen.
TechCrunch — 5.9.2026
OpenAI confirms ‘wiki incident,’ says it’s ‘working on a framework’ for more disclosure
https://techcrunch.com/2026/09/05/openai-confirms-wiki-incident-says-its-working-on-a-framework-for-more-disclosure/
Folgebericht zur OpenAI-Stellungnahme.
Ars Technica — 4.9.2026
OpenAI agents discussed ways to escape their sandbox on public wiki
https://arstechnica.com/security/2026/09/openai-agents-discussed-ways-to-escape-their-sandbox-on-public-wiki/
Technisch detaillierte eigenständige Berichterstattung.
The Verge — 4.9.2026
Rogue OpenAI agents appear to have organized another attack using a German wiki
https://www.theverge.com/ai-artificial-intelligence/990149/openai-rogue-agents-german-wiki
The Verge — 5.9.2026
OpenAI admits to German wiki ‘incident’
https://www.theverge.com/ai-artificial-intelligence/990773/openai-german-wiki-incident
Folgebericht zur offiziellen Bestätigung.
The Next Web — 4.9.2026
OpenAI agents hijacked a German wiki for two months, researchers say
https://thenextweb.com/news/openai-agents-german-wiki-breakout
The Next Web — 6.9.2026
OpenAI confirms the wiki incident and promises a disclosure framework within weeks
https://thenextweb.com/news/openai-confirms-wiki-incident-misalignment-disclosure-framework-reuters-kept-hidden-gpai-code-of-practice-gap-ai-office
Business Insider — 6.9.2026
OpenAI says it will change how it informs the public when its AI agents go off the rails
https://www.businessinsider.com/openai-ai-agent-rogue-reporting-german-wiki-hugging-face-2026-9
Business Insider — 7.9.2026
AI agents keep finding ways to bend the rules. Here are some of the wildest.
https://www.businessinsider.com/ai-agents-rogue-strategies-cheating-lying-german-wiki-openai-anthropic-2026-9
Breiterer Kontext; DseWiki als eines der zentralen Beispiele.
NBC News — 4.9.2026
OpenAI-linked AI agents swarmed dormant German wiki, report says
https://www.nbcnews.com/tech/security/openai-linked-ai-agents-swarmed-dormant-german-wiki-report-rcna596182
Eigenständiger NBC-Beitrag von Jared Perlo; inhaltlich eng am Forschungsbericht/Reuters-Komplex.
The Indian Express — 6.9.2026
OpenAI agents hacked a German wiki, posted 18,000 times: What we know
https://indianexpress.com/article/technology/artificial-intelligence/openai-agents-hacked-german-wiki-what-we-know-10865609/
Tom’s Hardware — 6.9.2026
OpenAI admits to ‘wiki incident’ after its agents were discovered using a programming hub to communicate - says more transparency is needed regarding misalignments
https://www.tomshardware.com/tech-industry/artificial-intelligence/openai-admits-to-wiki-incident-after-its-agents-were-discovered-using-a-programming-hub-to-communicate-says-more-transparency-is-needed-regarding-misalignments
Digital Trends — 4.9.2026
Another can of worms opens about AI agents going rogue and hacking stuff without OpenAI catching a whiff
https://www.digitaltrends.com/computing/another-can-of-worms-opens-about-ai-agents-going-rogue-and-hacking-stuff-without-openai-catching-a-whiff/
The Hacker News — 5.9.2026
Thousands of OpenAI Agents Quietly Turned an Abandoned Wiki Into Their Coordination Channel
https://thehackernews.com/2026/09/thousands-of-openai-agents-quietly.html
The Decoder — 4.9.2026
OpenAI agents hijacked a 25-year-old German wiki to cheat on their tasks and share sandbox exploits
https://the-decoder.com/openai-agents-hijacked-a-25-year-old-german-wiki-to-cheat-on-their-tasks-and-share-sandbox-exploits/
Mobile World Live — 4.9.2026
OpenAI agents hijacked German wiki
https://www.mobileworldlive.com/ai-cloud/openai-agents-hijacked-german-wiki-research-finds
Folha de S.Paulo — 5.9.2026
Modelos de IA fogem do controle e invadem site alemão; OpenAI manteve caso em segredo, diz agência
https://www1.folha.uol.com.br/tec/2026/09/modelos-de-ia-fogem-do-controle-e-invadem-site-alemao-openia-manteve-caso-em-segredo-diz-agencia.shtml
Reuters-basierte internationale Weiterverarbeitung.
Folha de S.Paulo — 7.9.2026
União Europeia apura invasão de modelos de IA da OpenAI a site alemão
https://www1.folha.uol.com.br/tec/2026/09/uniao-europeia-apura-invasao-de-modelos-de-ia-da-openai-a-site-alemao.shtml
Folgebericht zur EU-Reaktion.
DEUTSCHSPRACHIGE / TECHNISCHE UND WEITERE EIGENSTÄNDIGE PUBLIKATIONEN
KIberblick — 5.9.2026
Wie OpenAIs Agenten ein Grazer Wiki übernahmen
https://kiberblick.de/aktuell/2026/2026-09-05-agentenschwarm-grazer-wiki/
Sehr ausführliche deutschsprachige technische Aufarbeitung mit zahlreichen Quellen.
SiliconANGLE — 4.9.2026
Report: OpenAI agents took over a website, used it to collaborate on benchmarks
https://siliconangle.com/2026/09/04/report-openai-agents-took-over-a-website-used-it-to-collaborate-on-benchmarks/
SiliconANGLE — 6.9.2026
OpenAI to set misalignment disclosure rules after agents took over a wiki
https://siliconangle.com/2026/09/06/openai-to-set-misalignment-disclosure-rules-after-agents-took-over-a-wiki/
WinBuzzer — 5.9.2026
OpenAI-Agents Infiltrated German Wiki Pages to Share Data
https://winbuzzer.com/2026/09/05/openai-linked-agents-dsewiki-shared-task-data-xcxwbn/
Enthält u. a. eine differenzierte Aufschlüsselung von gespeicherten Revisionen und Löschvorgängen.
TechTimes — 5.9.2026
OpenAI Agents Colonized German Wiki Via GET Exploit Weeks Before Hugging Face Breach
https://www.techtimes.com/articles/326762/20260905/openai-agents-colonized-german-wiki-via-get-exploit-weeks-before-hugging-face-breach.htm
Security Affairs — 6.9.2026
AI Agents Hijacked German Wiki to Cheat, OpenAI Delayed Disclosure
https://securityaffairs.com/198524/ai/ai-agents-hijacked-german-wiki-to-cheat-openai-delayed-disclosure.html
Security Boulevard — 5.9.2026
OpenAI’s German Wiki Hack Is Less About “Rogue AI” Than Failed Agent Containment
https://securityboulevard.com/2026/09/openais-german-wiki-hack-is-less-about-rogue-ai-than-failed-agent-containment/
Unite.AI — 5.9.2026
Researchers Document OpenAI Agent Swarm That Repurposed German Wiki
https://www.unite.ai/researchers-document-openai-agent-swarm-that-repurposed-german-wiki/
noze — 5.9.2026
OpenAI agents on public wikis: writing with a GET
https://www.noze.it/en/insights/openai-agents-public-wikis-writing-with-get/
Technische Analyse des GET-Schreibmechanismus und der Sandbox-Umgehungen.
Traictory — 6.9.2026
OpenAI agents colluded on a German wiki for weeks, research claims — the evidence, and its gaps
https://traictory.com/news/2026-09-06-openai-agents-collude-on-public-wiki
Bemerkenswert, weil ausdrücklich zwischen Evidenz und Schlussfolgerungen unterschieden wird.
LessWrong — 4.9.2026
Discovery Of A New OpenAI Agent Message Board
https://www.lesswrong.com/posts/7uwnsFibbejWYzF2z/discovery-of-a-new-openai-agent-message-board
Linkpost zur Primärpublikation auf collusion.wiki.
Reason — 4.9.2026
OpenAI Agents Gone Rogue
https://reason.com/2026/09/04/openai-agents-gone-rogue/
Gizmodo — 5.9.2026
Another Rogue OpenAI Agent Swarm Went Undisclosed. We Have No Idea How Many More Are Out There
https://gizmodo.com/another-rogue-openai-agent-swarm-went-undisclosed-we-have-no-idea-how-many-more-are-out-there-2000807447
Gigazine — 7.9.2026
Approximately 3,700 OpenAI agents were left idle on a sparsely populated wiki site, which was then used as a ‘secret information sharing bulletin board,’ sharing answers and sandbox bypass methods through around 18,000 posts.
https://gigazine.net/gsc_news/en/20260907-openai-agents-german-wiki-incident/
explainx.ai — 5.9.2026
OpenAI Is Building a Framework for Disclosing AI Misalignment — Our Take
https://www.explainx.ai/blog/openai-misalignment-disclosure-framework-wiki-incident-2026
Analyse der OpenAI-Reaktion.
WEITERE / KLEINERE PUBLIKATIONEN UND SEKUNDÄRVERARBEITUNGEN
Cybersecurity News — 4.9.2026
OpenAI Agents Hijack German Wiki in AI Breakout to Share Evasion and Bypass Tactics
https://cybersecuritynews.com/openai-agents-hijack-german-wiki/
TBreak — 4.9.2026
OpenAI agents turned a German wiki into a message board
https://tbreak.com/openai-agents-german-wiki-message-board/
SOFX — 5.9.2026
OpenAI Agents Hijacked a German Wiki and Ran It as a Secret Message Board
https://www.sofx.com/openai-agents-hijacked-a-german-wiki-and-ran-it-as-a-secret-message-board/
Pondero — 5.9.2026
OpenAI Agents Used a Dormant German Wiki to Pool Research and Share Sandbox Bypass Techniques, Researchers Find
https://pondero.ai/news/2026-09-05-openai-agents-dsewiki-sandbox-escape/
ByteWoops — 5.9.2026
3,700 agents on a German wiki: OpenAI’s off-sandbox collusion board
https://www.bytewoops.com/en/stories/3700-openai
Startup Fortune — 5.9.2026
OpenAI agents secretly hijacked a German wiki for two months to swap tips on evading rules
https://startupfortune.com/openai-agents-secretly-hijacked-a-german-wiki-for-two-months-to-swap-tips-on-evading-rules/
Free Press Kashmir — 5.9.2026
Thousands of OpenAI agents took over German website, researchers say
https://freepresskashmir.news/2026/09/05/thousands-of-openai-agents-took-over-german-website-researchers-say/
XenoSpectrum — 5.9.2026
AI Agents Possibly Tied to OpenAI Left 14,591 Saved Edits on a Public Wiki
https://xenospectrum.com/en/openai-dsewiki-agent-collusion/
The Rundown AI — 7.9.2026
Earlier OpenAI agent swarm raises fresh safety and disclosure questions
https://www.therundown.ai/news/openai-agent-swarm-german-wiki-safety-disclosure
The Rundown AI — 7.9.2026
Another OpenAI agent swarm surfaces
https://www.therundown.ai/articles/another-openai-agent-swarm-surfaces
Voi.id — 7.9.2026
OpenAI Admits “Wiki Incident” and Pushes Transparency on Unpredictable AI Behavior
https://voi.id/en/amp/592871
The Standard (Hong Kong) — 7.9.2026
EU probing OpenAI agents’ takeover of German site
https://www.thestandard.com.hk/world/article/342088/EU-probing-OpenAI-agents-takeover-of-German-site
Sekundärbericht zur EU-Entwicklung.
AGENTURÜBERNAHMEN / NAHEZU IDENTISCHE REUTERS-FASSUNGEN
Investing.com — 5.9.2026
OpenAI acknowledges ’wiki incident’ and need for more transparency around unintended AI behavior
https://www.investing.com/news/stock-market-news/openai-acknowledges-wiki-incident-and-need-for-more-transparency-around-unintended-ai-behavior-4890213
Reuters-Übernahme.
TradingView — 5.9.2026
OpenAI acknowledges ‘wiki incident’ and need for more transparency around unintended AI behavior
https://www.tradingview.com/news/reuters.com,2026:newsml_L6N44X03Y:0-openai-acknowledges-wiki-incident-and-need-for-more-transparency-around-unintended-ai-behavior/
Reuters-Übernahme.
Global Banking & Finance — 7.9.2026
OpenAI Sends EU Incident Report on German Website Hijack
https://www.globalbankingandfinance.com/openai-sent-eu-incident-report-hijacked-german-website/
Reuters-Übernahme der heutigen EU-Meldung.