Courts may require lawyers to check AI-generated material before filing it, but Harvard lecturer and defence attorney Duncan Levin argues that questioning a lawyer about a false filing could raise a separate constitutional issue: the right against self-incrimination. He says that does not give lawyers a blanket right to refuse questions about AI use.
Watch Desk analysis
What happened
In an interview with Harvard Law Today, Levin discusses a possible clash between court efforts to verify AI-assisted filings and the Fifth Amendment. One proposal from the US Court of Appeals for the Tenth Circuit would require lawyers and people representing themselves to certify that AI-prepared filings had been checked by a human before submission.
Levin’s argument is that a certification could record what a lawyer represented, while a later question about whether they actually checked the material might raise a different issue. He says the privilege may apply if a truthful answer could expose someone to criminal prosecution, but not simply to avoid professional embarrassment, malpractice claims or ordinary sanctions. Whether it applies depends on the question and circumstances, and courts would decide.
Why it matters
The distinction is between asking people to stand behind what they file and compelling them to give potentially incriminating testimony about their conduct. Levin says courts can pursue accuracy and act on evidence they already have, while the Fifth Amendment may limit compelled answers in some cases.
That is an argument about how existing constitutional principles might apply, not a settled ruling on AI-specific court inquiries. Levin says the Supreme Court has not decided that question.
Our read
Human review of AI-generated filings is a sensible safeguard against fabricated cases and citations. The harder question is how courts enforce it without treating every inaccurate filing as evidence of a crime, or every lawyer’s silence as proof of wrongdoing. The law has not supplied a neat AI-era answer, and the interview is careful not to pretend otherwise.
What to watch
- Whether the Tenth Circuit adopts its proposed certification requirement.
- How courts distinguish mistakes or negligence from conduct that could create criminal exposure.
- Whether future rulings clarify how the Fifth Amendment applies to questions about AI-assisted filings.
Discussion spark: Should courts require lawyers to certify that they checked AI-generated material, even if that certification could later complicate questioning about a false filing?
Sources and evidence
- Source update (15 October 2026, 17:40 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.