Deepfakes may soon become a practical problem for international arbitration. Artificial intelligence (AI) has become increasingly integrated into arbitral procedure and decision-making, as discussed in a previous Aceris Law note on LaPaglia v. Valve. AI tools are frequently praised for their capacity to produce large volumes of text, images, audio and video within seconds. Yet their growing use has also exposed one of their most important limitations: their tendency to “hallucinate”, that is, to produce false, misleading, or entirely fabricated information while presenting it as reliable.
At first glance, such errors may appear manageable in legal proceedings. Fabricated legal authorities, however serious, can generally be detected through judges’ verification by assessing the accurate representation of references used. Indeed, in June 2026, a federal judge in Mississippi sanctioned lawyers on both sides after reading their briefs, which cited non-existent case law.[1]
Nevertheless, the difficulty is greater where falsification concerns the evidence itself. Deepfakes do not merely risk introducing inaccurate information into the record; they may also undermine the tribunal’s ability to determine whether apparently reliable evidence is genuine at all. An image, audio recording or video may appear authentic, while in fact depicting a person, statement or event that has been artificially generated or manipulated. The image below illustrates this problem: without further verification, can one know whether it is authentic, AI-generated or manipulated?
This is important in international arbitration, where the tribunal’s reasoning and final decision often depend heavily on documentary, audio and video evidence. If audiovisual material can be generated or manipulated in a manner that is difficult to detect, arbitral tribunals may increasingly be required to decide not only what weight to give to evidence, but also how to manage disputes about whether such evidence is authentic in the first place.
As Sejin Kim observed, deepfakes have greatly weakened the traditional assumption that what appears on screen is authentic.[2] Similar concerns have been expressed by Hany Farid in The New York Times, who warned that the unprecedented development of deepfake technology may be outpacing society’s ability to detect and respond to it.[3]
The issue is therefore not only whether deepfakes can be detected, but how arbitral tribunals should respond when the authenticity of audiovisual evidence is disputed. The procedural challenges that follow will also be addressed.
I. Deepfakes in International Arbitration as an Evidentiary and Procedural Risk
Deepfakes are audiovisual content generated or manipulated, through artificial intelligence, to alter a person’s appearance, voice, statements and conduct or to distort the representation of an event.[4]
There does not yet appear to be any publicly reported international arbitration case in which a party has submitted proven deepfake video evidence. The absence of such a case, however, should not suggest that the problem is theoretical. Arbitral tribunals have already had to deal with disputes concerning the authenticity, admissibility and reliability of recordings or other contested evidence. Thus, deepfakes do not create an entirely new evidentiary problem; rather, they intensify an existing one.
For instance, EDF (Services) Limited v. Republic of Romania (2009) illustrates that arbitral tribunals may refuse to rely on audiovisual evidence where there are unresolved doubts as to its authenticity. The tribunal declined to admit an audio tape and transcript produced as new evidence because their authenticity had not been established.[5]
Moreover, Libananco Holdings Co. Limited v. Republic of Turkey (2008) shows that tribunals may adopt procedural safeguards where the integrity of the arbitral process is threatened. In the case, the claimant alleged that privileged and confidential communications between the claimant, its counsel and potential witnesses had been intercepted. The tribunal treated those allegations as engaging fundamental principles of the ICSID arbitral process, including procedural fairness, confidentiality, legal privilege, and the parties’ right to present their case without interference.[6]
Furthermore, the 2023 decision of the Higher Regional Court of Cologne in OLG Cologne – 19 Sch 34/22 illustrates the concerns that may arise where a tribunal relies on a party-produced video recording to decide technically complex matters without appropriate expert assistance.[7] In that decision, the court considered whether reliance on a unilateral notarial video inspection of a production line, without adequate technical assessment, had violated the respondent’s right to be heard.
In sum, audiovisual evidence often raises questions of authenticity, lawful and fair collection, and technical assessment. Deepfakes make each of these questions more difficult, as they make manipulation easier, detection harder, and objections to authenticity more likely.
Challenging genuine evidence poses a serious danger to what Charlie Morgan and Michelle Penelope King describe in AI and Arbitration as “evidentiary integrity”.[8] As confidence in audiovisual evidence decreases, parties may be tempted to argue that authentic but unfavourable material has been altered by AI. This phenomenon has been called a “Liar’s Dividend” by legal scholars or even the “Deepfake Defence”.[9]
In arbitration, this risk has a specific procedural dimension. Tribunals often take a flexible approach to admissibility, leaving questions of reliability and probative value to be assessed later.[10] Thus, there is an initial presumption of authenticity of evidence. If challenged, the burden of proof rests on the party relying on the evidence.[11]
However, such a flexible approach may become harder to maintain in cases involving deepfakes. If tribunals admit audiovisual evidence too readily, they risk relying on manipulated material. If, however, they require expert verification whenever AI manipulation is alleged, parties may use authenticity objections to delay the proceedings, increase costs, or attempt to exclude harmful evidence at the admissibility stage.
Tribunals therefore need a middle ground. They should not treat audiovisual evidence as automatically reliable, but they should also be cautious about accepting unsupported allegations of AI manipulation. Challenging audiovisual evidence should require the identification of concrete reasons for doubt before the tribunal shifts the burden of authentication. Otherwise, deepfake objections may become a tactic for delaying proceedings or excluding damaging evidence.
II. The Existing Evidentiary Rules, Their Limits, and the Tribunal’s Gatekeeping Role
International arbitration has developed tools to assist tribunals on evidentiary matters, such as the IBA (International Bar Association) Rules on the Taking of Evidence in International Arbitration.
However, the IBA Rules are not strictly mandatory. Rather, they are designed to be adopted by arbitral tribunals and parties “in whole or in part”, in conjunction with the applicable procedural law and the circumstances of the case.[12] This flexibility is one of the strengths of international arbitration, but it may also create uncertainty when tribunals are faced with new technological risks such as deepfakes.
The IBA Rules assume that parties and legal representatives will act in “good faith”.[13] That assumption is central to the fairness of the arbitral process.[14] Yet, deepfakes may make tribunals unable to determine immediately whether a party has submitted fabricated evidence, or whether the opposing party has raised an authenticity objection in good faith.
Article 9(1) of the IBA Rules gives tribunals authority to decide on the “admissibility, relevance, materiality and weight” of evidence. However, the rules do not expressly identify AI-manipulated evidence as a separate ground for exclusion. In some circumstances, deepfake evidence may be addressed through existing categories, such as evidence “obtained illegally”.[15] Nevertheless, the absence of an express reference to AI manipulation suggests that existing evidentiary rules have not fully adapted to the specific risks posed by synthetic media.
A similar gap, indicated by Rachel Howie and Flavio Peter, may be observed in the Chartered Institute of Arbitrators’ Framework Guideline on the Use of Technology in International Arbitration, which regulates certain technological uses, including virtual reality for hearings, but does not provide a dedicated framework for deepfakes.[16]
Deepfakes also raise concerns beyond procedural admissibility. Because they may reproduce a person’s image or voice, they may interfere with personality rights.[17] This issue has already arisen in litigation involving AI-generated content. As explained by Hairuo Zhang and Yaxi Wang, Chinese courts have considered the unauthorised reproduction of a person’s voice.[18] The concern was not simply whether the content was accurate, but whether the person’s voice or image had been used without consent. Such cases also raise issues of consent, identity, privacy and personality rights.
Article 1(4) of the IBA Rules confirms the decisive role that tribunals play: they can interpret the rules set where a conflict arises regarding the IBA Rules’ meaning. The tribunal’s role is therefore not limited to admitting or excluding evidence; it also includes preserving the fairness and integrity of the proceedings.
* * *
In conclusion, in the deepfake context, the central procedural difficulty is to distinguish genuine authenticity concerns from tactical objections. A party challenging audiovisual evidence should therefore be required to identify concrete reasons based on definite indicators such as inconsistencies in metadata, unexplained editing, irregularities in the chain of custody, or other technical anomalies.
Technical tools and procedural measures such as hash verification, digital watermarking, chain-of-custody protocols and pre-hearing evidentiary reviews may help establish the provenance and integrity of audiovisual evidence, although they cannot always prove that the underlying content is genuine.[19]
[1] A. Polge, Aux États-Unis, un procès a été annulé par une juge après que les avocats des deux parties ont cité des jurisprudences totalement inventées par une IA, Science & Vie, 15 June 2026.
[2] S. Kim, Audiovisual Evidence in International Arbitration: Would ‘Seeing is Believing’ still work?, in C. González-Bueno (ed.), 40 under 40 International Arbitration (2021), pp. 211-232.
[3] E. Saslow, The World’s Leading Deepfake Expert No Longer Trusts His Own Eyes, The New York Times, 14 June 2026.
[4] LexisNexis, Deepfakes, https://www.lexisnexis.co.uk/legal/guidance/deepfakes (last accessed 25 June 2026).
[5] EDF (Services) Ltd. v. Romania, ICSID Case No. ARB/05/13, Award, 8 October 2009, ¶ 225.
[6] Libananco Holdings Co. Ltd. v. Republic of Turkey, ICSID Case No. ARB/06/8, Decision on Preliminary Issues, 23 June 2008.
[7] OLG Köln, Order of 8 May 2023 – 19 Sch 34/22.
[8] C. Morgan and M. P. King, A Human Story: AI, Arbitration and the Importance of Judgment, in S. Nappert and F. Carvalho Dias de Oliveira Silva (eds.), AI and Arbitration (2026).
[9] Ibid.; R. A. Delfino, The Deepfake Defense – Exploring the Limits of the Law and Ethical Norms in Protecting Legal Proceedings from Lying Lawyers, 84(5) Ohio St. L.J. 1067, p. 1068.
[10] N. Pitkowitz and M. Maurer, The Arbitrator and the Arbitration Procedure, The Vienna Proposition on Efficiency in Arbitration, in C. Klausegger et al. (eds.), Austrian Yearbook on International Arbitration 2026 (2026), p. 287.
[11] Kim, p. 225.
[12] 2020 IBA Rules on the Taking of Evidence in Intl. Arbitration, Preamble, para. 2.
[13] Id. Preamble, para. 3.
[14] Id. Preamble, para. 1.
[15] Id. Art. 9(3).
[16] R. Howie and F. Peter, Introduction to Evidence in International Arbitration, in C. Lotfi et al. (eds.), International Arbitration in Practice (2025), p. 230.
[17] D. Zhang, Integrating New Technologies in International Arbitration: Thoughts on the Future of Our Profession, in P. Schöldström and C. Danielsson (eds.), Stockholm Arbitration Yearbook Series, Volume 6 (2024), p. 108.
[18] H. Zhang et al., Annual Review on Intellectual Property Dispute Resolution in China (2025), in Commercial Dispute Resolution in China: An Annual Review and Preview (2025), p. 434.
[19] Morgan and King.
