ChatGPT
AI-generated text. This page was generated using artificial intelligence.
ChatGPT is a conversational AI service developed and operated by OpenAI. Its processing of personal data became the subject of a temporary restriction and investigation by Italy's data-protection authority in March 2023.[1]
Data-protection intervention in Italy
On March 31, 2023, the Garante announced an immediate temporary limitation on OpenAI's processing of Italian users' data. The authority identified concerns about notice, the legal basis for training-data processing, inaccurate personal information and age verification. These were the authority's stated grounds for intervention, not a description of an AI system independently making policy.[1]
On April 28, 2023, the authority reported that OpenAI had restored access in Italy after introducing changes. These included information notices, ways to object to training uses of personal data, and age-related registration measures. The Garante said its investigation would continue and identified further requested work. Reopening access did not amount to a comprehensive finding of compliance.[2]
General-purpose AI governance
The May 2024 International scientific report on the safety of advanced AI: interim report identified applications such as ChatGPT as a route through which general-purpose AI became familiar to the public. It described unusually strong public and policymaker interest and considered the opportunities and risks of systems capable of many tasks.[3]
That report supplies a documented policy connection, but not proof that ChatGPT alone caused any particular law to be adopted. This profile covers the system and selected regulatory history; OpenAI covers the organization's own policy proposals and institutional role.[3]
September 17, 2026: litigant responsibility
The Employment Appeal Tribunal considered a 300-page ChatGPT-generated argument in Hancox v Sutherland. It emphasized litigants’ own responsibility for checking accuracy and complying with procedural requirements; AI assistance is not generally prohibited.[4] See September 17 digest coverage.
Related articles
- Texas consumer-enforcement action concerning ChatGPT — Texas’s October 2026 petition alleges deceptive marketing and SCOPE Act violations and requests restrictions on minors’ access and new model releases. These are allegations and requested remedies, not court findings or orders.[5]
- Michael Murray — Research and teaching on AI-assisted legal practice and education.
- Cruz Barriga v. Sam’s West, Inc. — The October 9, 2026 protective order bars submitting defendant-produced confidential and attorneys’ eyes only discovery to open AI platforms, expressly including ChatGPT.[6]
- USA Today publisher copyright action concerning training and outputs.
- Sam Altman
- AI consciousness — Historical foundations, scientific evidence and policy debate over artificial experience and welfare.
- British Columbia v. Altman — The province and a school district’s September 2026 lawsuit.
- September 22, 2026 — British Columbia’s lawsuit over Tumbler Ridge
- OpenAI 2026 Model Hacking Incidents — Incident chronology, overlapping reports and September 16, 2026 disclosures.
- X Corp. and X.AI LLC v. Apple Inc. et al. — AI antitrust litigation
- Beus Gilbert PLLC v Brigham Young University — sanctions for unverified AI-assisted briefing, September 9, 2026.[7]
- State v. Sandoval / In re Stephen D. Aarons — AI-assisted briefing sanctions
- September 10, 2026 nonparty-discovery ruling in the Meta copyright cases
- OpenAI
- Policy
- United States policy on catastrophic AI risk
- European Union policy on catastrophic AI risk
October 2026 teen-safety assessment
Common Sense Media called for adult-only access on October 7, 2026 after testing teen safeguards and reporting parental-alert and crisis-referral gaps.[8] OpenAI disputes the testing methodology.[9] See October 7 digest coverage.
Advertising and conversation context (October 2026)
On October 21, 2026, OpenAI described ChatGPT’s advertising system as a separate service using different models and code from answer generation. It said ad selection uses conversation context and, when users choose personalized ads, available nonsensitive information from past conversations.[10]
OpenAI says ads are labeled, conversations remain private from advertisers and customer data are not sold.[11] These company-described safeguards distinguish use of conversation context to select ads from influence on answer generation. See OpenAI’s advertising and conversation-context controls and October 21 news coverage.
References
- ↑ 1.0 1.1 Garante, Artificial intelligence: stop to ChatGPT by the Italian SA, March 31, 2023. Accessed September 12, 2026.
- ↑ Garante, OpenAI reinstates service in Italy, April 28, 2023. Accessed September 12, 2026.
- ↑ 3.0 3.1 International scientific report on the safety of advanced AI: interim report, May 2024, executive summary and introduction. Accessed September 12, 2026.
- ↑ Hancox v Sutherland and Others [2026] EAT 139, September 17, 2026, cover, summary and paragraphs 1–3, 31–37, 39–40.
- ↑ Texas’s Original Petition, dated October 23, 2026, pp. 59–65, paragraphs 190–209.
- ↑ District of Nevada, Cruz Barriga v. Sam’s West, Inc., ECF 13, October 9, 2026, paragraph 19 and approval at pages 10 and 12.
- ↑ U.S. District Court for the District of Utah, Memorandum Decision and Order Imposing Sanctions, ECF 385, September 9, 2026.
- ↑ Common Sense Media, October 7, 2026.
- ↑ Axios via Yahoo, October 7, 2026.
- ↑ Ritendra Datta and Karthik Gopalratnam, OpenAI, “How we built ChatGPT’s novel ad stack for user trust and relevance,” October 21, 2026, complete written account.
- ↑ OpenAI, “ChatGPT Ads expands to 26 new countries as over 100,000 advertisers have used the platform,” October 21, 2026, complete written announcement.