[W]e reiterate to the Court Reporter that this tribunal depends on transcripts that faithfully mirror the proceedings as they occurred.
In the Thursday opinion addressing Williams v. State, the ruling was authored by Indiana Court of Appeals Judge Felix and was joined by Chief Judge Tavitas and Judge Bradford:
Even though the adage that “[t]rial records are rarely if ever perfect,” as stated in Ben-Yisrayl v. State, 753 N.E.2d 649, 662 (Ind. 2001), is acknowledged, the transcript in this matter does not come close to ideal.
The transcript is fraught with a range of faults. A multitude of typographical mistakes alter the meaning of testimony, questions, or objections. See, e.g., Tr. Vol. II at 137:18, 144:10, 147:10; Tr. Vol. III at 6:13. In several instances, the names of witnesses and trial attorneys are recorded inaccurately. Tr. Vol. II at 220:5; Tr. Vol. III at 142:15–20, 143:15, 162:4–5.
At one point, a motion—likely filed by the State—is attributed to the trial court. Tr. Vol. II at 107–08. At another point, an objection—likely raised by Williams—is attributed to the Bailiff. Tr. Vol. II at 177:15. And elsewhere, the State’s closing argument is attributed to the trial court. Tr. Vol. III at 228:1.
These errors, along with others not enumerated here, made Williams’s appeal harder to review but did not prevent the court from conducting a meaningful examination. Nevertheless, we remind the Court Reporter that this court relies on transcripts that faithfully reflect the proceedings that were transcribed. See Ind. Appellate Rule 28(B) (“The Court Reporter shall certify the Transcript is correct.”); see, e.g., Orr v. State, 277 N.E.3d 1262, No. 25A-CR-1786 (Ind. Ct. App. Mar. 6, 2026) (mem.) (identifying similar errors), trans. denied, 280 N.E.3d 249 (Ind. 2026).
Given the kinds of flaws identified, it appears that a generative artificial intelligence system may have assisted in compiling the transcript. While AI can enhance efficiency and serve as a useful tool across many professions, those who use such technology bear the responsibility to proofread and verify the accuracy of the output it generates.
Rob Freund (@RobertFreundLaw) brought this to light on Thursday, and 404 Media (Samantha Cole) covered the development.
