Indiana Court Flags AI Errors in Official Transcript, Warns Reporter
Indiana court officially identified AI-caused errors in a court reporter's transcript.
Why it matters: This case highlights serious risks of relying on AI for critical legal documents without sufficient human verification. Legal professionals must maintain strict oversight to ensure accuracy in proceedings.
- Indiana Judge Paul Felix found AI-related mistakes in a drug-overdose case transcript, including misattributed statements.
- The Indiana Court of Appeals reminded the court reporter that transcripts must be true, accurate records.
- Similar AI errors have appeared in the UK, Georgia, and Connecticut, involving fake legal citations and case law.
- Judges and courts emphasize the need for proofreading and verification when using AI-generated legal documents.
In a recent ruling, Indiana Court of Appeals Judge Paul Felix identified errors likely caused by AI in an official transcript filed by a court reporter in a drug-overdose case. The transcript contained misattributions, such as a State motion incorrectly credited to the trial court and a defendant's objection wrongly attributed to the Bailiff.
The court formally reminded the reporter that it depends on transcripts to be accurate representations of the proceedings. Judge Felix stressed, "It is incumbent upon those using such systems to proofread and ensure the accuracy of the generated product." This highlights the critical need for human oversight in legal transcription, especially as AI tools become more common.
This issue is not isolated. The UK Crown Prosecution Service admitted to presenting fabricated case law in a High Court extradition appeal, attributing errors to AI but underscoring that failure to verify authorities was the root cause according to CPS.
Similarly, a Georgia Court of Appeals panel vacated part of a ruling after a trial court used a proposed order with AI-generated fake legal citations warned by the appeals court. Connecticut attorney Ian G. Gottlieb also apologized to the state Supreme Court for filings containing AI-generated errors, including invented case citations, as detailed by the Connecticut Post. Justice Andrew J. McDonald warned that relying on AI as a shortcut "debas[es] the professionalism of the practice."
These incidents across multiple jurisdictions underscore the dangers of unverified AI use in legal documentation. They reinforce the necessity of rigorous human review to maintain the integrity and trustworthiness of court records and filings.
By the numbers:
- 4 jurisdictions — Indiana, UK, Georgia, Connecticut with reported AI transcription or citation errors
- Multiple instances — AI errors include misattributions, fabricated citations, and fake case law
- 3 judges and courts — publicly warn of risks and emphasize human proofreading
Yes, but: While AI promises efficiency, human oversight remains essential to prevent serious inaccuracies in legal records.
What's next: Indiana and Connecticut courts may impose sanctions or issue new guidelines on AI use in legal transcription and filings.