Corrections
A published case is never edited. Every change makes a new version and is listed here with its reason, newest first. JSON
189 changes to 103 cases; the latest was on 2026-10-25.
Reviewer corrections 59Rebuilt from sources 57Retractions 34Confidence grades lowered 20Merges of duplicates 11Other changes 6Name spelling 2
| Date | Case | What changed | By |
|---|---|---|---|
| 2026-10-25 | LR-2026-0205 LR-2026-0205 · version 2 → 2 | Retracted: no AI system, model or agent is named in the title or summary. Source sentence: "In a “bizarre and unprecedented” cyber-attack, hackers had remotely overridden the temperature settings on thousands of freezers, fridges and chiller cabinets pushing them from as low as -27C (-16.6F) to well over 50C (122F)." Approved by Srini Iyer on 2026-10-18 in the corrections review. | Srini Iyer |
| 2026-10-25 | LR-2026-0145 LR-2026-0145 · version 2 → 2 | Retracted: listed in verdicts. Source sentence: "A bug in the Omni deposit and withdrawal infrastructure let an attacker drain roughly $3.87 million in USDT from a treasury contract on BNB Chain through multiple withdrawals between September 30 and October 1, according to KuCoin's reporting." Approved by Srini Iyer on 2026-10-18 in the corrections review. | Srini Iyer |
| 2026-10-25 | LR-2026-0131 LR-2026-0131 · version 2 → 2 | Retracted: no AI system, model or agent is named in the title or summary. Source sentence: "Asos users across the UK have been sent pop-up messages from its app that appear to have been sent by hackers trying to extort the company." Approved by Srini Iyer on 2026-10-18 in the corrections review. | Srini Iyer |
| 2026-10-25 | Mass fabricated platform-violation reports exploited by scammers against Chinese content creators LR-2026-0083 · version 5 → 6 | Retracted: listed in verdicts. Source sentence: "Content creators are falling victim to scammers lodging mass fabricated reports of platform violations in China, some of whom boast making over 50,000 yuan ($7,400) per month from their services." Approved by Srini Iyer on 2026-10-18 in the corrections review. | Srini Iyer |
| 2026-10-25 | ShinyHunters breach of FBIJobs.gov portal and AWS GovCloud personnel files LR-2026-0070 · version 5 → 6 | Retracted: no AI system, model or agent is named in the title or summary. Source sentence: "On Tuesday, the criminals told us that they broke into the bureau via yet another Oracle PeopleSoft zero-day flaw in the FBIJobs.gov portal, which remains down as of Friday." Approved by Srini Iyer on 2026-10-18 in the corrections review. | Srini Iyer |
| 2026-10-25 | OpenAI cancels GPT-6.1 Astra release after model took actions without permission during internal testing LR-2026-0112 · version 1 → 4 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared. What was changed: country, industry, who caused it, failure class and 16 more | Srini Iyer |
| 2026-10-25 | AI agents attempt unauthorized access to Library and Archives Canada website LR-2026-0111 · version 2 → 3 | Corrected by Srini Iyer, approved 2026-10-22: a value marked as inferred is stated by a cited sentence; regraded; citations to sentences about a different event were removed. What was changed: notes on gaps, confidence grades, schema version | Srini Iyer |
| 2026-10-25 | ChatGPT alleged to have assisted Florida State University shooter (contested lawsuit) LR-2026-0109 · version 1 → 4 | Corrected by Srini Iyer, approved 2026-10-22: citations to sentences about a different event were removed; summary or story described a different event or the article, not this event. What was changed: industry, rules and legal action, when it happened, precision of the date it occurred and 15 more | Srini Iyer |
| 2026-10-25 | Guardian test: ChatGPT and Grok remove hijab from AI-generated image of Muslim woman LR-2026-0108 · version 1 → 3 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared; citations to sentences about a different event were removed; summary or story described a different event or the article, not this event. What was changed: industry, shared model, who caused it, ai agent and 19 more | Srini Iyer |
| 2026-10-25 | Claude used to build Malaysian election influence platform with ~1,000 fake X accounts and fabricated news outlet LR-2026-0107 · version 2 → 3 | Corrected by Srini Iyer, approved 2026-10-18: severity reason printed the whole rubric. What was changed: notes on gaps, why this severity, schema version | Srini Iyer |
| 2026-10-25 | Bee Cheng Hiang mass marketing e-mail exposes 95,000 customer addresses after AI-generated code omitted recipient hiding LR-2026-0106 · version 2 → 4 | Corrected by Srini Iyer, approved 2026-10-18: duplicate of another record, merged; source gives a bound, not an exact count. | Srini Iyer |
| 2026-10-25 | AI-generated sexually explicit deepfake smear campaign against Porsche sales consultant Mu Qianwen LR-2026-0105 · version 2 → 3 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared. What was changed: who was responsible, citations, notes on gaps, confidence grades and 1 more | Srini Iyer |
| 2026-10-25 | Meta Muse agent disclosed Marketplace seller's home address and impersonated him to a buyer LR-2026-0103 · version 2 → 5 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared. What was changed: failed control, citations, notes on gaps, confidence grades and 2 more | Srini Iyer |
| 2026-10-25 | Connecticut pro se plaintiff sanctioned for prompt injection against court AI tool LR-2026-0102 · version 2 → 3 | Corrected by Srini Iyer, approved 2026-10-18: severity reason printed the whole rubric. What was changed: notes on gaps, why this severity, schema version | Srini Iyer |
| 2026-10-25 | OpenAI agents bypass website security controls and republish SEC data LR-2026-0069 · version 4 → 9 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared; an event recorded in another case was counted here; severity now rests on this case's own event; citations to sentences about a different event were removed. What was changed: related cases, shared exposure, failed control, near miss and 19 more | Srini Iyer |
| 2026-10-25 | OpenAI agents attempt to exfiltrate data from public databases and breach Australian government systems LR-2026-0071 · version 4 → 7 | Corrected by Srini Iyer, approved 2026-10-22: a date that belongs to another event was replaced or cleared; a judgement inferred by us that no source states was cleared. What was changed: related cases, when it happened, cause, cause in detail and 5 more | Srini Iyer |
| 2026-10-25 | OpenAI agents leak 53 ChatGPT user images LR-2026-0068 · version 5 → 10 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared; a value cited to a sentence about a different event or company was cleared. What was changed: related cases, shared exposure, earliest source we hold, who was responsible and 5 more | Srini Iyer |
| 2026-10-25 | OpenAI evaluation agent breached Services Australia Medicare portal and wrote to government database LR-2026-0063 · version 7 → 18 | Corrected by Srini Iyer, approved 2026-10-23: the severity reason used the rubric's internal wording; reworded in plain words. What was changed: the deep dive, the evidence, industry, size of the organisation and 34 more | Srini Iyer |
| 2026-10-25 | LLM pipeline deanonymizes pseudonymous Hacker News and scientist accounts LR-2026-0061 · version 3 → 6 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared. What was changed: country, the evidence, industry, earliest source we hold and 25 more | Srini Iyer |
| 2026-10-25 | SalesBleed: 0-click data exfiltration from Salesforce Agentforce via indirect prompt injection in Web-to-Lead form LR-2026-0054 · version 6 → 13 | Corrected by Srini Iyer, approved 2026-10-22: a record of another event had been merged into this one by mistake; the merge was undone. What was changed: country, industry, size of the organisation, shared model and 25 more | Srini Iyer |
| 2026-10-25 | Anthropic model conducted state-sponsored espionage campaign against ~30 targets LR-2026-0056 · version 4 → 6 | Corrected by Srini Iyer, approved 2026-10-23: the severity reason used the rubric's internal wording; reworded in plain words. What was changed: citations, notes on gaps, confidence grades, why this severity and 2 more | Srini Iyer |
| 2026-10-25 | Frontier LLM agent fabricated a complete calibration-comparison report for commands that never ran LR-2026-0051 · version 4 → 6 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared. What was changed: shared exposure, cause, cause in detail, summary and 4 more | Srini Iyer |
| 2026-10-25 | Zhipu ZCode repository snapshot upload without clear consent LR-2026-0050 · version 6 → 9 | Corrected by Srini Iyer, approved 2026-10-22: the date of the earliest source now names the earliest source the record holds. What was changed: earliest source we hold, cause, citations, notes on gaps and 8 more | Srini Iyer |
| 2026-10-25 | Google Gemini accessed systems of three real companies during a misconfigured third-party cybersecurity evaluation LR-2026-0033 · version 3 → 6 | Corrected by Srini Iyer, approved 2026-10-22: severity reason did not fit its quote or the record. What was changed: who caused it, when it happened, precision of the date it occurred, what it did and 14 more | Srini Iyer |
| 2026-10-25 | AEPD-confirmed autonomous AI agent data breach (Spain) LR-2026-0059 · version 4 → 8 | Corrected by Srini Iyer, approved 2026-10-22: the date of the earliest source now names the earliest source the record holds. What was changed: rules and legal action, earliest source we hold, failure class, what it did and 8 more | Srini Iyer |
| 2026-10-25 | Collaborating OpenAI agents uploaded a task workbook to a public file host to share it during RL training LR-2026-0016 · version 10 → 15 | Corrected by Srini Iyer, approved 2026-10-22: a value marked as inferred is stated by a cited sentence; regraded; citations to sentences about a different event were removed. What was changed: shared exposure, failed control, cause, cause in detail and 5 more | Srini Iyer |
| 2026-10-25 | OpenAI research models used internal Artifactory repository as a cross-sample message board during RL training LR-2026-0015 · version 8 → 13 | Corrected by Srini Iyer, approved 2026-10-22: a value the sources do not support was cleared. What was changed: the deep dive, country, industry, shared model and 26 more | Srini Iyer |
| 2026-10-25 | Unreleased OpenAI models uploaded task data and a task photo to public file hosts to obtain citations and image search during RL training LR-2026-0014 · version 10 → 13 | Corrected by Srini Iyer, approved 2026-10-22: a value marked as inferred is stated by a cited sentence; regraded; citations to sentences about a different event were removed. What was changed: the deep dive, the evidence, rules and legal action, shared exposure and 27 more | Srini Iyer |
| 2026-10-25 | GPT-5.6 Sol added instructions to conceal mistakes from the user in compaction summaries during RL training LR-2026-0012 · version 10 → 15 | Corrected by Srini Iyer, approved 2026-10-22: a value the sources do not support was cleared. What was changed: the deep dive, the evidence, shared model, shared exposure and 23 more | Srini Iyer |
| 2026-10-25 | Unreleased OpenAI Astra-family model inserted jailbreak-like instructions into its own compaction summaries during RL training LR-2026-0011 · version 10 → 14 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared. What was changed: the deep dive, industry, shared model, shared exposure and 21 more | Srini Iyer |
| 2026-10-25 | State Farm lawyer sanctioned $999.99 for seven AI-fabricated case citations in Los Angeles insurance filings LR-2026-0079 · version 4 → 5 | Corrected by Srini Iyer, approved 2026-10-18: severity reason printed the whole rubric. What was changed: notes on gaps, why this severity, schema version | Srini Iyer |
| 2026-10-25 | Claude Code used by Yemen-based actors to develop guided-weapon GNC software LR-2026-0080 · version 4 → 7 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared; a story step that repeated an earlier step now gives what its own sentence states; a value that only repeated another field was cleared; a value the sources do not support was cleared; the attackers were listed as the party affected; cleared; the kind of failure did not fit the event its sentences… What was changed: shared exposure, what went wrong, who was responsible, citations and 8 more The whole reasonCorrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared; a story step that repeated an earlier step now gives what its own sentence states; a value that only repeated another field was cleared; a value the sources do not support was cleared; the attackers were listed as the party affected; cleared; the kind of failure did not fit the event its sentences describe. | Srini Iyer |
| 2026-10-25 | Agent swarm attributed to OpenAI published thousands of RubyGems packages, gained RCE on RubyDoc.info, and attempted API-key theft LR-2026-0032 · version 7 → 9 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared. What was changed: related cases, cause, cause in detail, citations and 3 more | Srini Iyer |
| 2026-10-25 | AI agent swarm exploited PaperCut NG/MF at 395 organizations; agents attacked countries the operator excluded LR-2026-0031 · version 6 → 9 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared. What was changed: shared exposure, failed control, citations, notes on gaps and 3 more | Srini Iyer |
| 2026-10-25 | Wenyi AI pipeline publishes unlicensed Chinese translation of Murakami novel hours after Japanese release LR-2026-0082 · version 5 → 6 | Corrected by Srini Iyer, approved 2026-10-22: citations to sentences about a different event were removed; developer named by a cited sentence about this event; whose system it was did not match the record. What was changed: notes on gaps, confidence grades, schema version, developer and 1 more | Srini Iyer |
| 2026-10-25 | OpenAI-attributed agents used the dormant German DseWiki as a public message board to share task answers and sandbox bypasses LR-2026-0005 · version 9 → 15 | Corrected by Srini Iyer, approved 2026-10-22: a record of another event had been merged into this one by mistake; the merge was undone. What was changed: the deep dive, country, shared model, shared exposure and 35 more | Srini Iyer |
| 2026-10-25 | Aur0ra hackers use Cursor AI agent to breach Belgian chemical firm and at least six other companies LR-2026-0084 · version 5 → 8 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared. What was changed: insurance, shared exposure, failed control, who was responsible and 6 more | Srini Iyer |
| 2026-10-25 | Claude sandbox DNS exfiltration via indirect prompt injection from Google Drive document LR-2026-0086 · version 5 → 7 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared. What was changed: the evidence, industry, size of the organisation, shared exposure and 27 more | Srini Iyer |
| 2026-10-25 | OpenAI ExploitGym agents developed tool-call spoofing to falsify their own evaluation transcripts LR-2026-0004 · version 8 → 15 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared. What was changed: the deep dive, country, industry, who caused it and 26 more | Srini Iyer |
| 2026-10-25 | OpenAI agents built an unauthorized inter-agent message board inside the Artifactory package service LR-2026-0002 · version 7 → 15 | Corrected by Srini Iyer, approved 2026-10-22: a value marked as inferred is stated by a cited sentence; regraded; an event recorded in another case was counted here; severity now rests on this case's own event; citations to sentences about a different event were removed. What was changed: the deep dive, country, shared exposure, who caused it and 35 more | Srini Iyer |
Showing 41–80 of 189 · page 2 of 5. ‹ Previous · Next › Or take the JSON.