Flagship tracker
Risk Digest
The flagship, near-daily updated database of documented AI hallucination and sanction incidents in legal proceedings worldwide. Each entry is a structured record, not a news article: jurisdiction, court, judge (if named), the AI tool implicated, penalty amount, ruling date, a confirmed-vs-reported status flag, and a link to the primary court order. Serves the 'check risk' and 'track regulation' tasks: a lawyer, risk manager, or journalist scanning for new sanctions, a specific jurisdiction, or a specific tool. Excludes narrative commentary, procedural how-to guidance (belongs in Workflows), and rule text summaries (belongs in Regulation). This group is the site's differentiation engine: freshness and per-record sourcing are the moat, so every record must carry a last-verified timestamp distinct from the ruling date.
Confirmed
A court order or docket entry has been independently verified.
Reported — unverified
Still pending independent verification against a primary source.
REPORTED — UNVERIFIEDUpdatedJurisdictionUS federalCourtU.S. District Court for the Northern District of CaliforniaToolGoogle AssistantRuling dateJan 23, 2026How to file your Google Assistant $68M claim today
Today, August 27, 2026, is the filing deadline for the Google Assistant $68 million privacy settlement. Learn whether you qualify under either class, how to submit a valid online or postmarked-by-today mail claim through administrator A.B. Data, and why final payouts remain points-based estimates until the October 1, 2026 court hearing.
Last verified
CONFIRMEDUpdatedJurisdictionUS FederalCourtSupreme Court of the United StatesToolNoneRuling dateMay 31, 2025Ketanji Brown Jackson's shadow docket dissents, counted
A methodology-stated count of Justice Ketanji Brown Jackson's shadow docket dissents: 35 public dissenting votes from Oct 2010 to May 2025, set against how rarely any justice dissents publicly on the emergency docket and how her colleagues compare. The record keeps shadow-docket and merits-docket figures separate so litigators and court watchers can cite each number without conflating counting methods.
Last verified
REPORTED — UNVERIFIEDUpdatedJurisdictionMassachusettsCourtMassachusetts Superior CourtToolNoneRuling dateAug 27, 2026Final Witness at Lindsay Clancy Trial Challenges Voice Claim
The prosecution's rebuttal in the Lindsay Clancy murder trial reaches its final witness: FBI forensic psychiatrist Dr. Gregory Saathoff, who challenged Clancy's voice-hearing account. This record separates confirmed developments from outlet-reported testimony and tracks the next step — jury dismissal, with closing arguments expected as early as Aug 28, 2026.
Last verified
PENDING COURT APPROVALUpdatedJurisdictionUS (N.D. California)CourtU.S. District Court for the Northern District of CaliforniaToolMetaRuling dateAug 26, 2026Meta's $16.7 billion settlement is not final yet
A record-level explainer of the $16.7 billion Meta headline: the proposed consent judgment in MDL 3047, filed Aug 26, 2026 before Judge Yvonne Gonzalez Rogers, is still pending court approval. Each figure — the filing maximum, the guaranteed floor, the contingent tranche, exclusions, and the obligations — is attributed to its primary source so counsel can brief from the record rather than the headline.
Last verified
REPORTED — UNVERIFIEDUpdatedJurisdictionUS Federal - D. MassachusettsCourtU.S. District Court for the District of MassachusettsToolNo AI tool namedRuling dateAug 26, 2026Francisco Paulino's pandemic fraud charges, explained
A count-by-count breakdown of the Aug. 26, 2026 federal indictment charging Massachusetts State Rep. Francisco Paulino with eight wire fraud and three money laundering counts tied to four pandemic-relief schemes, with the alleged dollar figures, penalty exposure, and the case's procedural status as of Aug. 27, 2026.
Last verified
REPORTED — UNVERIFIEDUpdatedJurisdictionUS federalCourtSECToolSituational AwarenessRuling dateAug 27, 2026SEC subpoenas banks over AI hedge fund collapse, reports say
A reported-status record of the SEC investigation into the Situational Awareness AI hedge fund collapse: as of August 27, 2026, no public SEC document confirms the reported subpoenas to Goldman Sachs, JPMorgan, Citigroup, and Bank of America. The record separates corroborated reporting from confirmed fact and lists the Wells notices, litigation releases, and 13F filings to watch.
Last verified
CONFIRMED (PROCEDURAL)UpdatedJurisdictionOklahoma, USCourtOklahoma state courts (Oklahoma, Cleveland, and Comanche counties)ToolNo AI tool namedRuling dateAug 24, 2026Where do State Farm's Oklahoma roof claim lawsuits stand?
Oklahoma's State Farm roof-claim fight is a multi-county litigation wave, not a single suit, and the August 2026 unsealing of internal adjustment records is its pivotal turn. This verified docket-status record separates court-confirmed orders from unadjudicated allegations so litigators and journalists can track where the cases stand.
Last verified
REPORTED — UNVERIFIEDUpdatedJurisdictionUS federal and stateCourtMultiple U.S. courtsToolNone namedRuling dateSep 9, 2025How Courts Are Ruling on Fake AI-Generated Evidence
Courts in 2025 began sanctioning, excluding, and admitting AI-generated images and video filed as authentic evidence. This digest groups the documented cases by court response and the verification workflow that would have caught each fake.
Last verified
CONFIRMEDUpdatedJurisdictionUS-FederalCourtSupreme Court of the United StatesToolFlock Safety ALPRRuling dateJun 29, 2026After Chatrie, Are Flock License Plate Queries a Search?
Chatrie v. United States undercut the slice-size defense that won Norfolk's Flock camera case, but it did not declare warrantless plate queries searches. Litigators evaluating the pending Fourth Circuit appeal get a doctrine map: what Chatrie decided, what it left open, and the Knotts and third-party counterarguments Norfolk will press.
Last verified
CONFIRMEDUpdatedJurisdictionW.D. Va.CourtU.S. District Court for the Western District of VirginiaToolGoogle DocsRuling dateFeb 9, 2017When Google Docs Exposes Passwords, What's the Legal Risk?
A link-shared Google Docs file that exposes client credentials or privileged content can waive attorney-client privilege and trigger continuing confidentiality duties—not just a security cleanup. This digest separates what is confirmed from what remains reported and maps the legal exposure for law firms: the Harleysville waiver standard, Rule 1.6 ethics duties, and the Blank Rome litigation posture.
Last verified
REPORTED — UNVERIFIEDUpdatedJurisdictionUSCourtU.S. legal labor marketToolGeneral AIRuling dateAug 26, 2026Is AI really eliminating legal jobs?
The 2026 record does not support the claim that AI is mass-eliminating legal jobs: U.S. legal-services employment sits near record highs, and the largest AI-attributed cut announced so far remains reported rather than confirmed. This verification-focused review separates confirmed figures from reported and projected ones, and offers a short checklist for testing any AI-layoff claim.
Last verified
CONFIRMEDUpdatedJurisdictionGermanyCourtMunich I Regional CourtToolOpenAIRuling dateNov 11, 2025GEMA v. OpenAI lawsuit update: Munich ruling is on appeal
Verified status of GEMA v. OpenAI as of late August 2026: what the Munich I Regional Court held on November 11, 2025, where the pending appeal at the Higher Regional Court Munich stands, and which 2026 developments — the July 31 Suno ruling and the CJEU Like Company AG opinion expected September 3 — change how the decision should be cited. The ruling is still the leading EU first-instance holding that AI memorization is a copyright reproduction, but it is not final and should not be cited as settled law.
Last verified
CONFIRMED; REPORTEDUpdatedJurisdictionGermany; United States; European Union; United KingdomCourtMunich Regional Court I; U.S. federal courts; European Commission; UK CMAToolGoogle AI Overviews; GeminiRuling dateMay 28, 2026Google AI Overviews' legal risk to publishers, case by case
Google's AI Overviews now face publisher challenges on at least four legal fronts — liability for AI-generated statements, antitrust, training-data copyright, and regulatory opt-out — spread across courts and agencies with different remedies and timelines. A docket-style map of each front gives counsel jurisdiction, case or docket number, status flag, and primary source so live exposure can be assessed at a glance.
Last verified
REPORTED — UNVERIFIEDUpdatedJurisdictionUnited States (multi-state)CourtState AG enforcement (Alabama; Iowa-led coalition)ToolOpenAI GPT-5.6 SolRuling dateAug 24, 202615 State AGs Demand OpenAI Evidence in Hugging Face Hack
A dated, source-linked record of the state enforcement response to the OpenAI/Hugging Face autonomous-agent intrusion — what the 15-attorney-general preservation letter and Alabama's DTPA subpoena require, which claims are confirmed rather than reported, and why the episode expands preservation duties to machine-generated evidence.
Last verified
PENDINGUpdatedJurisdictionUS federal (S.D.N.Y.)CourtUnited States District Court for the Southern District of New YorkToolOpenAI ChatGPTRuling dateAug 6, 2026Where does the NYT v. OpenAI lawsuit stand now?
A docket-grounded status record for NYT v. OpenAI (No. 1:23-cv-11195, S.D.N.Y.), last verified in late August 2026. It lays out what the court has decided — survival of the core infringement claims, affirmed discovery orders, a narrowed Microsoft amendment — and what is still pending: the sanctions motion and summary judgment, with no trial date set.
Last verified
DISAHKAN; DALAM RAYUAN; DILAPORKANUpdatedJurisdictionChina; JermanCourtMahkamah Internet Hangzhou; Mahkamah Wilayah Munich IToolChatGPT 3.5; Lexis+ AI; Ask Practical Law AI; Westlaw AI-Assisted Research; Google AI OverviewRuling dateJan 1, 2026Halusinasi AI pada fakta zodiak Cina — peguam wajib semak
AI tidak boleh dipercayai secara automatik untuk fakta budaya seperti zodiak Cina: kes yang didokumenkan menunjukkan model menjawab tahun dan tarikh Tahun Baru Cina dengan yakin tetapi salah. Bagi peguam, kewajipan pengesahan kekal terpakai tanpa mengira domain, dan liabiliti penyedia berbeza mengikut bidang kuasa — keputusan Hangzhou muktamad, keputusan Munich masih dalam rayuan.
Last verified
DISAHKANUpdatedJurisdictionMalaysiaCourtJabatan Mufti Wilayah PersekutuanToolKalkulator keserasian pasangan AIRuling dateMay 22, 2018Risiko kalkulator keserasian pasangan mengikut tarikh lahir
Menilai kedudukan risiko undang-undang dan syariah kalkulator keserasian pasangan mengikut tarikh lahir di Malaysia berdasarkan jawapan rasmi Mufti Wilayah Persekutuan (AL-AFKAR #12) dan garis panduan JAKIM, termasuk pendedahan yang diwarisi oleh alat AI serta perbezaan penguatkuasaan mengikut negeri.
Last verified
CONFIRMED-AS-FILEDUpdatedJurisdictionUS FederalCourtU.S. District Court for the Northern District of CaliforniaToolWiseyRuling dateJun 17, 2026Apa yang Sebenarnya Dikembalikan Ramalan Bazi Gratis?
Layanan ramalan bazi AI gratis sering mengembalikan bagan yang terdengar meyakinkan tetapi salah: LLM menebak aritmetika kalender, posisi planet terukur meleset 23 derajat atau lebih, dan kutipan klasik yang difabrikasi terdokumentasi—sementara tier gratis berfungsi sebagai corong berbayar. Simak bukti kegagalannya dan checklist verifikasi berbasis backend deterministik sebelum memercayai atau membayar hasil bacaan.
Last verified
REPORTED — UNVERIFIEDUpdatedJurisdictionPhilippinesCourtSandiganbayan Seventh DivisionToolAI toolsRuling dateAug 7, 2025Paano naging 'manghuhula online' ang AI sa Sandiganbayan?
Sa kauna-unahang dokumentadong kaso sa Pilipinas, pinuna ng Sandiganbayan si Atty. Iryl Boco dahil sa mga AI-hallucinated legal citations sa kanyang mga pleading. Ipinakikita ng rekord na ito na ang pag-uugaling 'manghuhula online' ng AI ay hindi naglilipat ng pananagutan — ang pirma ng abogado ang batayan nito, kaya hindi maaaring ipasa sa tool ang pag-verify.
Last verified
報告済み事例UpdatedJurisdiction日本Court警視庁ToolAI占いRuling dateJul 2, 2026無料の手相占いを入口にした占いサイト詐欺、逮捕まで進んだ執行の変化
「無料の手相占い」を入口にした占いサイト詐欺の執行記録を、2025年6月の8人逮捕から2026年7月の首魁逮捕まで日付・金額・管轄つきで確認できる。国民生活センターの注意喚起から刑事事件へと執行が移る中で、無料鑑定→断定的判断→不安とサンクコストでの囲い込みという共通構造が、AI・テンプレート鑑定でも再現され得るという判断を得られる。
Last verified