AI Enforcement Tracker
Courts, plaintiffs and regulators act against AI more often every year. We count three kinds of events: court decisions about AI errors in filings, private lawsuits about AI, and actions by regulators and attorneys general.
Events per quarter
AI enforcement events per quarter, 2019 to now, by stream. The scale is logarithmic: the steps from 1 to 10, 10 to 100 and 100 to 1,000 have the same height. The dashed gold line is the fitted trend.
- All events
- Courts acting on AI errors in filings
- Private AI lawsuits filed
- Regulator and attorney-general actions
- Trend: doubles every 14.0 months
Events by stream
| Stream | 2019 | 2023 | 2024 | 2025 | 2026 to date | Doubling time, months (95% range) |
|---|---|---|---|---|---|---|
| Courts acting on AI errors in filings | 0 | 16 | 61 | 865 | 1,320 | – |
| Private AI lawsuits filed | 14 | 71 | 86 | 169 | 256 | 20.5 (14.0–38.7) |
| Regulator and attorney-general actions | 10 | 51 | 92 | 143 | 150 | 17.5 (13.5–24.7) |
| All events | 24 | 138 | 239 | 1,177 | 1,726 | 14.0 (9.9–23.9) |
Without the court stream, enforcement doubles about every 19.4 months (15.7–25.6). Court decisions on AI errors grew from 16 in 2023 to 865 in 2025. We do not fit them on their own: they start in 2023, so there are too few years. At this pace, 2026 ends with about 2,060 events.
Latest events
| Date | Stream | Where | What | Source |
|---|---|---|---|---|
| 2026-11-26 | Court: AI error | USA: C.D. California | Orlando Chechitelli v. Etiwanda School District, et al. AI errors in a filing by a lawyer. Outcome: Bar Referral. | Source |
| 2026-10-29 | Regulator | United States | SEC v. BG Wealth Sharing Ltd / DSJ Exchange Pty Ltd (lawsuit or charge): $130M scheme promising returns from 'SI-generated signals' (SI = Super Intelligence, the… | Source |
| 2026-10-29 | Lawsuit | Cal. Superior Court, San Francisco | Feinberg v. OpenAI et al. Claim: harm from an AI product. | Source |
| 2026-10-29 | Court: AI error | USA: CA Indiana | Nabors v. Nabors. AI errors in a filing by a self-represented litigant. Outcome: Admonishment. | Source |
| 2026-10-29 | Court: AI error | USA: CA Texas | In the Estate of John Raburn Ford, Deceased. AI errors in a filing by a lawyer. | Source |
| 2026-10-28 | Regulator | Florida | Florida AG v. Rengstorff Consulting Ltd. (A2E AI platform) (lawsuit or charge): FDUTPA suit alleging the A2E generative AI platform lets users make nonconsensual nude… | Source |
| 2026-10-28 | Court: AI error | USA: SC New York | PC-59 Doe v Jewish Bd. of Family & Children's Servs., Inc. AI errors in a filing by a lawyer. Outcome: Monetary Sanction. | Source |
| 2026-10-28 | Court: AI error | USA: N.D. California | Farhat v. Cutera, Inc. AI errors in a filing by a lawyer. Outcome: Admonishment. | Source |
| 2026-10-28 | Court: AI error | USA: 10th Cir. CA | Contreras Ramirez v. Blanche. AI errors in a filing by a self-represented litigant. Outcome: Warning. | Source |
| 2026-10-28 | Court: AI error | USA: D. Wyoming | Clifford v. Neidle et al. AI errors in a filing by a self-represented litigant. Outcome: Brief Struck; Order to Show Cause. | Source |
| 2026-10-27 | Lawsuit | Cal. Superior Court, San Francisco (pleaded as JCCP 5431… | Wan et al. v. OpenAI (Bowen Wan). Claim: harm from an AI product. | Source |
| 2026-10-27 | Court: AI error | USA: D. Maryland | Quzack-White et al. v. Avanath Largo Center, LLC. AI errors in a filing by a lawyer. Outcome: Brief Struck; Order to Show Cause. | Source |
| 2026-10-27 | Court: AI error | USA: D. D.C. | Matiella v. Murdock Street, LLC. AI errors in a filing by a lawyer. | Source |
| 2026-10-27 | Court: AI error | USA: D. Colorado | Rivers v. Town of Hudson et al. AI errors in a filing by a self-represented litigant. Outcome: Order to Explain. | Source |
| 2026-10-27 | Court: AI error | USA: CA New York (1d) | Abuzaid v Almayouf. AI errors in a filing by a self-represented litigant. Outcome: Warning. | Source |
| 2026-10-26 | Court: AI error | USA: W.D. Virginia | Fridenstine v. Midkiff. AI errors in a filing by a lawyer. Outcome: Warning. | Source |
| 2026-10-26 | Court: AI error | USA: CA Illinois (2d) | McCray v. Lee. AI errors in a filing by a self-represented litigant. | Source |
| 2026-10-23 | Regulator | Texas | Texas AG v. OpenAI; Sam Altman (lawsuit or charge): DTPA (and SCOPE Act) suit in Collin County alleging OpenAI misrepresented ChatGPT safety and failed to protect users… | Source |
| 2026-10-23 | Lawsuit | N.D. Cal. | Steven Checchia and Stacy Penning v. OpenAI OpCo, LLC. Claim: privacy or biometrics. | Source |
| 2026-10-23 | Lawsuit | N.D. Tex. | SpaceX Super Intelligence LLC v. Martinez. Claim: deepfakes. | Source |
| 2026-10-23 | Lawsuit | N.D. Cal | ABG EPE IP, LLC et al. v. Anthropic PBC. Claim: copyright or IP. | Source |
| 2026-10-23 | Lawsuit | U.S. District Court for the Northern District of… | Checchia v. OpenAI (ad-tracking wiretapping). Claim: privacy or biometrics. | Source |
| 2026-10-23 | Court: AI error | USA: C.D. California | United States of America et al. v. Los Angeles Film School LLC et al. AI errors in a filing by a lawyer. Outcome: Monetary Sanction; CLE; Bar Referral. | Source |
| 2026-10-23 | Court: AI error | South Africa: Labour Court of South Africa | South African Post Office SOC Ltd (In Business Rescue) and Others v Commissioner Menzi Fakude N.O. and Others. AI errors in a filing by a judge and a lawyer. Outcome… | Source |
| 2026-10-23 | Court: AI error | USA: CA Arizona | Lewis v. Cameron-Nichols. AI errors in a filing by a self-represented litigant. Outcome: Adverse Costs Order. | Source |
The newest 25 of 3,434 events. Sources add older cases late, so a new row can have an old date. The CSV has every event.
Money
Courts and regulators imposed or approved $6.98B over AI from 2019 to now. This counts fines, sanctions, costs orders, settlements and verdicts with a known amount, in US dollars at the rate of the decision date. Three cases are 55% of the total:
- Bartz v. Anthropic PBC, 2026: $1.50B (court settlement)
- Texas AG v. Meta Platforms, 2024: $1.40B (regulator settlement)
- AP (Dutch DPA) v. Uber B.V., 2026: $965M (regulator fine, under appeal)
Overturned decisions are not counted. Some amounts are under appeal or have only preliminary approval. Settlements with secret terms are not visible.
Method
- One event is one court decision about AI errors in a filing, one private lawsuit about AI (counted on its filing date), or one public action by a regulator, attorney general or prosecutor (counted on its announcement date). We count the three streams separately and add them.
- Courts acting on AI errors. Every decision in Damien Charlotin's AI Hallucination Cases database: a court or tribunal found that a filing had invented cases, false quotes or similar AI errors.
- Private AI lawsuits. The union of the GWU Database of AI Litigation (DAIL) and ailawsuittracker.com, private plaintiffs only, with duplicates removed by case name and date. We add curated lists of AI copyright suits, chatbot harm suits and AI defamation suits.
- Regulator and attorney-general actions. A hand-checked list of 501 actions. Each one links to the authority's own announcement. An action counts when the announcement makes AI central to the violation or the remedy.
- Doubling time. Least squares of ln(events per year) on year, for full years 2019–2025. Doubling time = 12 × ln 2 ÷ slope. The 95% range uses the t-distribution. A year enters the fit only when all three streams cover it fully.
- Weekly update. Each week a script reads the Charlotin database, DAIL and ailawsuittracker.com again, rebuilds the counts, fits the trend again and publishes this page.
Data dates
- Courts acting on AI errors in filings: data to 2 November 2026 (database update of 2026-11-02).
- Private AI lawsuits filed: data to 2 November 2026 (DAIL read 2026-11-02, ailawsuittracker.com read 2026-11-02; curated lists to 2026-10-30).
- Regulator and attorney-general actions: data to 30 October 2026 (curated by hand to 2026-10-30; no automatic update).
Limits
- Lower bounds. Every source misses cases. A court must write about an AI error before it appears. Small state-court suits, police cases and most non-English sources are thin.
- Late additions. The sources add old cases late. Recent months are always too low. Older years were filled in later, which can make growth look faster.
- The court stream is most of the count. It was 73% of events in 2025. It measures AI errors that lawyers and litigants file, as much as AI failure. A court sanctions the person who filed, not the AI vendor.
- The regulator list is not automatic. No public register lists all AI enforcement, so we built the list by hand. It stops on 30 October 2026. The curated lawsuit lists also stop on that date. Until someone updates them, the latest year is too low for these streams.
- Money is lumpy. A few cases make most of the total. Face- and voice-recognition cases count as AI.
- No merge across streams. A lawsuit and a later attorney-general action on the same conduct are two events.
Sources and credit
- Damien Charlotin, AI Hallucination Cases database (HEC Paris), CSV export, CC BY 4.0.
- GW Law Ethical Tech Initiative, Database of AI Litigation (DAIL).
- AI Lawsuit Tracker (ailawsuittracker.com).
- AI copyright suits merged from ChatGPT Is Eating the World, BakerHostetler, McKool Smith, the Copyright Alliance, DAIL and ailawsuittracker.com, with dates checked on CourtListener.
- Regulator actions: the announcements of the FTC, SEC, DOJ, FCC, US state attorneys general, the European Commission, EU data protection authorities, the ICO, Ofcom, the CAC and authorities in Korea, Australia, Canada, Brazil and other countries. One link per event in the CSV.
- Exchange rates: European Central Bank reference rates.
Juniper Research analysis. Part of Juniper Research. Questions or a missing case: [email protected].