Boersma v. Davenport (Oregon 2026): $500 Sanction for AI-Fabricated Cases in Appellate Brief — Despite Being Warned in Prior Case | Advocate Prakhar

Case at a Glance

Background

Heather Boersma filed a pro se appeal from a limited judgment awarding defendants $735.00 for her failure to attend a deposition. During its review of the opening brief, the Oregon Court of Appeals identified several citations to completely fabricated cases and numerous instances where Boersma cited real cases for propositions they did not support. The court issued a show cause order asking why the brief should not be stricken and why monetary sanctions should not be imposed for submitting nonexistent case law.

In her response, Boersma claimed she conducted research in good faith despite limited access to legal research tools, did not intend to submit fabricated authority, and had been subjected to “heightened scrutiny for minor or correctable citation issues.” She pointed to a separate simultaneous appeal as evidence of her good faith — in that other appeal, opposing counsel had notified her of citation problems, and she had filed a motion for leave to file an amended brief acknowledging the errors.

The critical finding came from that separate appeal history: at the time Boersma filed the brief in this case, she was already aware from the other appeal that her legal research and writing process was producing fabricated authority. She took no steps to ensure the brief in this case was free of the same problems. The court found this “prior experience regarding citation errors” aggravated rather than mitigated her conduct in the present case.

The AI Issue

The Oregon Court of Appeals applied ORCP 17 C(1), which (through ORAP 1.40(4)) applies to appellate courts and requires any party who signs a filed document to certify that its contents are based on “reasonable knowledge, information and belief, formed after the making of such inquiry as is reasonable under the circumstances.” Self-represented parties must comply with court rules to the same extent as any other litigant. The fabricated citations, combined with Boersma’s awareness of identical problems in a simultaneous appeal, meant the court found sanctions were warranted despite her general good-faith claims.

What the Court Decided

• Filing completely fabricated cases and citing real cases for unsupported propositions in an appellate brief violates ORCP 17 C(1)’s reasonable-inquiry certification requirement.
• Self-represented parties must inform themselves of and comply with court rules as any other litigant — no relaxation applies to the citation-accuracy obligation.
• Knowledge of identical citation problems in a simultaneously filed brief in another case creates an obligation to audit all simultaneously filed briefs before those briefs reach the court — failure to do so is an independently aggravating factor.
• The court imposed a $500 monetary sanction payable to the Appellate Court Services Division of the Oregon Judicial Department.
• The appeal was also dismissed on a separate threshold issue: the limited judgment from which Boersma appealed was not a valid limited judgment and therefore was not appealable.

“Plaintiff apparently took no steps to ensure that her opening brief in this case did not contain the same issues to which she had been alerted in the separate appeal.”

— Oregon Court of Appeals, Boersma v. Davenport, May 6, 2026

The India Angle

Indian Law Equivalent

Under the Supreme Court of India’s inherent jurisdiction and the High Courts’ powers under Article 215 of the Constitution, courts can impose costs for frivolous proceedings and misrepresentations. In Ramjas Foundation v. Union of India (2010) 14 SCC 38, the Supreme Court imposed costs on parties who had misled the court on factual matters. An advocate or litigant who files fabricated citations in one proceeding and then takes no corrective steps in simultaneous proceedings before the same or related courts would face enhanced costs or contempt findings based on the prior-knowledge factor emphasized in Boersma.

Bar Council Rules

BCI Rules, Chapter II, Rule 9 requires advocates not to do anything that tends to mislead the court. The Boersma aggravating factor — knowledge of citation problems in one case, failure to apply that knowledge to correct simultaneously pending cases — would be particularly serious for Indian advocates, who often maintain multiple pending matters simultaneously. A lawyer who knows from one case that their AI research process is producing hallucinated citations and takes no corrective steps for other pending cases would face disciplinary proceedings under the BCI Rules for each affected case.

Practical Advice for Indian Advocates

• If you discover citation errors in one matter, immediately treat all other simultaneously pending matters as presumptively affected and conduct a citation audit of their filings before the next hearing date.
• When filing multiple briefs in overlapping timeframes using AI assistance, batch-verify all citations before any of the filings are submitted — problems in one brief typically indicate a systemic workflow issue that affects all briefs.
• Maintain a practice management log documenting when citation issues were identified and corrected across all matters, so that if a court inquires about your awareness of prior issues, you can demonstrate responsive action.

Quick Takeaways

• Awareness of citation problems in one case requires auditing all simultaneously filed briefs.
• $500 sanction payable to the court system — not opposing party — signals institutional harm focus.
• Good-faith claims are undermined when the filer had specific, actionable prior notice of the problem.

Deep Dive: The Cross-Case Knowledge Standard

Boersma v. Davenport introduces what can be called the cross-case knowledge standard for AI hallucination liability. The standard asks: did the filer know, at the time of filing, that their legal research and drafting process was prone to producing AI hallucinations? If yes, did they take reasonable steps to address that known risk before filing? Boersma knew — she had been specifically told in a simultaneous appeal that her brief contained citation errors — and she did nothing. That asymmetry between knowledge and action is what the court found sanctionable.

This standard creates an expanding zone of accountability for repeat AI users. The first time a practitioner discovers that their AI research tool has produced hallucinated citations, their liability is low: courts tend to treat it as an honest first mistake. But from that moment forward, the practitioner has actual knowledge that their AI use creates citation risks. Every subsequent brief filed without verification now carries heightened exposure because the practitioner cannot honestly claim ignorance of the risk.

The cross-case knowledge standard also has implications for law firms and legal teams. If a junior associate in a firm discovers an AI hallucination in one brief, does that knowledge impute to other members of the firm who are using the same AI tools for other briefs? While courts have not yet addressed this firm-level imputation question directly, the Boersma reasoning suggests that knowledge of a tool’s hallucination tendencies should trigger firm-wide verification protocols, not just individual-matter corrections.

For Indian advocates who use AI research tools across a wide practice portfolio, Boersma is a strong signal to implement systematic verification protocols that apply across all matters simultaneously rather than addressing citation problems on a case-by-case basis as they are discovered. The moment an AI hallucination is found in any matter, the appropriate response is a firm-wide audit of all pending filings that used the same tool or process — not a targeted fix only for the matter where the problem was identified.