1. What the document is: "OpenAI Government User Data Request Policy" — explains how law enforcement and government agencies can request user data or content removal from OpenAI. Outside this wiki's core focus (AI training and data-licensing risk); included here for completeness.
What this wiki found — complete, every page
Presence facts, not verdicts — each flagged term links to its definition and the exact clause on this page. Absence of a badge means the term isn't currently flagged here, not that the page is risk-free.
ChatGPT – Law Enforcement Guidelines
Dataset: GenGA (Generative AI Governance Archive) — 11 AI services, 2025–present
GenGA (Generative AI Governance Archive) is this wiki’s Generative-AI dataset: raw policy snapshots captured by the Open Terms Archive
genai-euproject across 11 GenAI providers (ChatGPT, Claude.ai, DeepSeek, Google Generative AI Services, Le Chat, Llama API, Meta AI, Microsoft Copilot, Perplexity, Qwen Chat, xAI). Unlike PGAv2, GenGA has no pre-tagged risk scores — all risk scoring and keyword tagging on these pages is LLM-assigned by direct reading, spanning 2025–present.
Note: This page contains documented policy clauses. Risk assessment is qualitative and context-dependent. For analysis of patterns across platforms, see:
Source: sources/GenGA/ChatGPT/Law Enforcement Guidelines/ (raw Markdown captures, Open Terms Archive genai-eu dataset — no pre-tagged JSONL exists for this source)
Last updated: 2026-02-18
Note on methodology: No pre-computed
risk_score/keywordfields exist for this source. Keyword tags on this page are LLM-assigned. Treat as first-pass analysis, not externally verified ground truth. This wiki’s risk-scoring system was retired project-wide on 2026-06-21 (seemethodology.md§4) — no page, GenGA included, computes or displays a numeric risk score.
Overview
1. What the document is: “OpenAI Government User Data Request Policy” — explains how law enforcement and government agencies can request user data or content removal from OpenAI. Outside this wiki’s core focus (AI training and data-licensing risk); included here for completeness.
2. Input rights: Not addressed. This is a government-request procedural policy, not a user-facing license document.
3. Output restrictions: Not addressed.
4. Non-explicit predatory clauses: None. The document is procedural — it covers what legal process is required, preservation periods, emergency disclosure, and user notice.
5. Regulatory references: References specific US law (18 U.S.C. § 2703(f), MLAT/18 U.S.C. § 2523) and Irish/EU legal frameworks for cross-border requests, plus general “international human rights laws, principles, and norms.” No EU AI Act or GDPR reference.
6. Regional variation: Yes, explicitly two-track: requests to OpenAI US (governed by US law — subpoenas, warrants, etc.) versus requests to OpenAI Ireland Limited (EEA/Switzerland users, governed by Irish law), each with its own preservation and emergency-disclosure procedures.
7. Key risk to users: Low, and not really within this wiki’s core risk scope. Worth noting: OpenAI “reserve[s] the right to seek reimbursement for costs associated with responding to law enforcement data requests” (§VII), and content-removal decisions rest on OpenAI’s own discretionary assessment of “the potential impact of the demand on the human rights of impacted users” (§VI), rather than a fixed legal standard.
No AI-training, data-licensing, or output-restriction clauses found — outside this wiki’s core risk categories, consistent with how Imprint/Brand Guidelines doc types are also treated.
Regulatory & Research Context
Davidson et al. (2026), in their comparative study of regulatory gray areas across OpenAI’s and four other providers’ Terms, describe a “post-API” world in which “researcher-specific access has been reduced, shifted toward ‘pay-to-play’ models, or simply removed” — a dynamic this document does not itself address, but one that contextualizes why OpenAI’s discretionary, self-assessed standard for content-removal requests (§VI’s “potential impact… on the human rights of impacted users”) sits alongside, rather than within, any externally-imposed access regime. Edwards et al. (2025)‘s “platformisation paradigm” — providers positioning themselves as neutral intermediaries while retaining discretionary control — is visible here in OpenAI’s right to “seek reimbursement for costs” and its own non-fixed legal standard for evaluating government data requests, rather than a court-imposed or statutory test.
Changes Summary
| Date | What changed |
|---|---|
| 2026-02-18 | Baseline and only version captured in this dataset (effective 1 January 2026, “v2025.12”). |
Version History
2026-02-18
- Explicit AI clause: NO
- Non-explicit predatory: NO
- Flagged keywords: (none)
No AI-training, data-licensing, or output-restriction clauses found in this document.