Net Good IndexSubmit a correction

Harm Ledger · verified · Computer Science

An Oklahoma judge acknowledges ChatGPT produced fictitious citations in a ruling

On 9 September 2026 Reuters reported that Stephens County District Judge Lawrence Wheeler told investigators he used ChatGPT for research and that an order he drafted contained two citations to cases that do not exist.

9 Sep 2026Tier 2 Significant HarmMethodology 0.1

Current score

−0.14

3 base · Significant Harm (tier 2 of 5, 3 pts)
× 0.3500 attribution · Material acceleration
× 0.8500 evidence · Multiple independent confirmations
× 0.9000 realization · Realized outcome
× 0.3500 durability
Event-level product before credit split: 0.28

False citations in a judicial order are significant institutional harm (tier 2). The judge used ChatGPT as a research aid (0.35). Journalism plus a prosecutor letter quoting the judge’s admission (0.85). Realization is a filed order with fictitious cases. Residual harm depends on correction and any discipline.

What happened

A prosecutor’s 17 August letter to the state attorney general, reviewed by Reuters, said Wheeler acknowledged the false citations. The same judicial-misconduct complaint also alleged the judge secretly listened to jury deliberations; prosecutors said the AI-citation issue did not support criminal prosecution. Harm is to the integrity of a state-court ruling. No population injury is documented.

Model attribution

−0.14

GPT

Used by the judge for research; generated two nonexistent case citations that entered a drafted order.

The judge admitted ChatGPT research produced the false citations; he issued the order.

Attribution 0.3500 · Credit share 50% · OpenAI

Claims

  • Judge Wheeler acknowledged that a ruling he drafted contained two ChatGPT-generated citations to cases that do not exist.

    outcome · supported

Sources

independent sources

Revision history

  • 13 Sep 2026 · 0.00 → 0.14

    Imported events under methodology 0.1.