By Leslie King O’Neal
What are “Deepfakes?”
“’Deepfakes’ are videos or images where a person’s face, body or voice has been digitally altered so they appear to be someone else, typically used maliciously or to spread false information.”[i] The rapid development of generative AI tools in the past few years made it much easier for almost anyone to create highly realistic (but false) photographs and videos. A quick Google search shows numerous tools available to create “deepfakes.”[ii]
Judge Dismisses Case for Plaintiffs’ Use of “Deepfake” Evidence
Recently, a California judge issued “terminating sanctions” (dismissal) of pro se plaintiffs’ claims for filing false AI-generated evidence in support of their motion for summary judgment.[iii] The court found plaintiffs violated the CA civil procedure code by submitting false evidence. Therefore, she dismissed the case with prejudice. The order is included below because it includes links to the AI-generated deepfake videos, photographs and text threads, which readers can review for themselves.
AI-Generated Witness Videos Were Deepfakes
Specifically, the court found that submitted witness videos were AI-generated by comparing them with another video of the same witness, stating: “While the ‘person’ depicted in exhibits 6A and 6C bears a passing resemblance to the person depicted in exhibit 36, they are not the same person. The accent, cadence, volume, word choice, pauses, gestures, and facial expression, among other characteristics, of the person depicted in exhibit 36 are vastly different from those demonstrated by the “persons” depicted in exhibits 6A and 6C.”
Plaintiffs Submitted Altered Photographs
Plaintiffs submitted photographs allegedly taken by their Ring camera. The court found these photographs had been altered. “The lighting, contrast, color, and sharpness of the man depicted in these pictures compared with the lighting, contrast, color, and sharpness of the background shows that the man was stitched into the photograph taken by the Ring camera.” Also, the background was in black and white, while the man was shown in color. Also, plaintiffs claimed the evidence “was captured on an Apple iPhone 6 Plus using its rear camera and running iOS 12.5.5.” However, the court noted, “Apple did not introduce Apple Intelligence until iOS18 and required an iPhone 16, iPhone 15 Pro or iPhone 15 Pro Max.”
Plaintiffs Submitted AI-Generated Text Threads
The court also found Plaintiffs’ text thread exhibits (allegedly from Instagram and Zoom chats) were AI-generated or materially altered because the menu style at the bottom of the threads was the same on both threads. Also, the font style, spacing, and sizing appeared “off.”[iv] The court listed several other exhibits it found “suspicious” but stated it “does not have the time, funding, or technical expertise to determine the authenticity of Plaintiffs’ statements or conduct a forensic analysis of the suspect evidentiary submissions.”
Court Considers Appropriate Sanctions
Noting that, “the use of deepfakes in a case significantly undermines the Court’s ability to administer justice, significantly erodes the public’s confidence in the judicial system, and significantly burdens under-resourced and overworked courts with the time-consuming task of assessing whether evidence presented to it during pretrial proceedings was a deepfake,” the court determined plaintiffs’ conduct warranted more severe sanctions than merely a monetary penalty. This was not merely a case of fictitious citations, but was a deliberate attempt to perpetrate a fraud on the court.
No Criminal Referral
The court determined that, although this conduct could be perjury or forgery that could subject plaintiffs to criminal prosecution, referral for prosecution was both too severe a penalty and not sufficiently remedial. Such a penalty would not address the harm done in this case, but could have wider repercussions for the plaintiffs’ lives going forward.
Evidence and Issue Sanctions Inadequate
The court found evidence and issue sanctions would not be an adequate deterrent for future use of false AI-generated evidence. While such sanctions would exclude some false evidence, there might be other fabricated evidence that neither the court nor opposing counsel could detect.
Court Finds Terminating Sanction Appropriate
Considering proportionality and the need to deter future use of false AI-generated evidence, the court determined that terminating the case was the appropriate sanction here. Consequently, the court dismissed the entire action with prejudice.
Courts & Arbitrators Grapple with Technology and Generative AI
Nearly every day brings new reports of cases involving AI “hallucinations” such as fake cases, phony facts, and inaccurate quotations.[v] Judges have imposed different types of sanctions on lawyers, law firms and clients for court filings with these issues– ranging from monetary fines to reporting lawyers to bar associations to dismissing cases with prejudice.[vi] So far these sanctions have not eliminated the problem. Under current arbitration rules, arbitrators lack much authority to sanction lawyers, witnesses or parties for arbitration filings with “hallucinations.” See post: Can Arbitrators Sanction Attorneys for AI Hallucinations?
Few Rules or Standards for Courts or Arbitrators re: AI Use
Although lawyers, consultants, clients, judges and arbitrators are using AI tools more frequently, there are few available standards or guidance on this issue. Some states have adopted laws or amended their evidence rules to cover AI use.[vii] Two organizations have published standards for AI use in arbitration.[viii] Currently the best guidance is to discuss AI use at a preliminary conference and include the parties’ agreement in the initial scheduling order.[ix]
Sanctions Under Current Arbitration Rules
Under AAA Rules, arbitrators may limit a party’s participation in a proceeding or may issue adverse rulings as sanctions. However, AAA rules do not allow arbitrators to issue default awards as sanctions.[x] If parties fail to follow JAMS Rules or arbitrator orders, arbitrators may award fees and costs and to make adverse inferences or, in extreme cases, to determine issues adversely to a party as sanctions. As noted in a prior post, including AI use requirements in preliminary orders gives arbitrators more authority to sanction misuse if necessary.[xi]
Takeaways
- Attorneys, arbitrators, clients, and experts are using AI tools more frequently.
- Discussing AI use and disclosure at the beginning of arbitration proceedings is a good practice.
- Attorneys and arbitrators should be aware of the potential for “deepfake evidence” in arbitration.
[i] Oxford English Dictionary (December 2024).
[ii] E.g. the author’s Google search for “creating deepfakes with ai” gave these results: “Ai Deep Fake Video Generator,” “Make Compelling Deepfake Videos,” Deepfake AI—Free Guide,” “Deepfakes Web| Make Your Own Deepfake [Online App].”
[iv] The court also found the metadata Plaintiffs submitted was unreliable. Further, the court found plaintiffs’ explanations were not credible.
[v] There is a database of cases involving AI hallucinated case citations. See Damien Charlotin, “AI Hallucination Cases” https://www.damiencharlotin.com/hallucinations/
[vi] Sanctions Order, Frankie Johnson v. Jefferson Dunn, et al. (Case No: 2:21-cv-1701-AMM, D.Ct. N.D. AL July 23, 2025) https://docs.justia.com/cases/federal/district-courts/alabama/alndce/2:2021cv01701/179677/204; Order (imposing sanctions), Byoplanet Int’l, Inc. v. Peter Johnsson, et al., (Case No: 0:25-cv-60630-LEIBOWITZ, D.Ct. S. D. FL July 17, 2025) https://caselaw.findlaw.com/court/us-dis-crt-sd-flo/117513089.html; Order to Show Cause (imposing sanctions), Eric Coomer Ph.D. v. Michael J. Lindell, et al. (Case No: 1:22-cv-01129 NYw SBP (D.Ct.Col. July 7, 2025)https://www.courtlistener.com/docket/63296393/309/coomer-v-lindell
[vii] See LA Rev. Stat. §14:73.14 (2024) https://law.justia.com/codes/louisiana/revised-statutes/title-14/rs-14-73-14/; Cal. Civ. Code—CIV §3344, amended by SB 683 (October 10, 2025).
[viii] Silicon Valley Arbitration and Mediation AI Guidelines, https://svamc.org/wp-content/uploads/SVAMC-AI-Guidelines-First-Edition.pdf; Chartered Institute of Arbitrators (CIARB) Guideline, Guidance 18 https://www.ciarb.org/resources/professional-practice-guidelines/international-arbitration/
[ix] AAA has published standards for AI use in ADR, www.adr.org/media/sx2mjcdj/aaai_standards_for_ai_in_adr_.pdf
JAMS has published special rules for AI disputes,https://www.jamsadr.com/artificial-intelligence-disputes-clause-and-rules
[x][x] Rule 61, AAA Construction Industry Arbitration Rules and Mediation Procedureshttps://www.adr.org/media/kydkdsn1/construction-industry-arbitration-rules-and-mediation-procedures-2024.pdf
[xi] https://theconstructionadrtoolbox.com/2025/07/can-arbitrators-sanction-attorneys-for-ai-hallucinations/