2025 in Review: Technology

2025 was marked by the continued integration and scrutiny of technology in arbitration. As artificial intelligence (“AI”) and other digital tools moved from novelty to normalised use, the arbitration community grappled with their potential to reshape practice — from procedural efficiency to the very nature of decision-making. At the same time, this evolution unfolded against ongoing debates over risk, regulation, and ethical boundaries.

Using Technology in Arbitration: Improvement or Reform?

Discussions on how the use of technology could affect the arbitration process remained a hot topic in 2025. Some contributions focused on whether technology might support more fundamental change. One commentator explored how the consensus on automation in the UNCITRAL Model Law on Automated Contracting could serve as a basis for automated, and eventually autonomous, arbitration. Another revisited how decentralised decision-making enabled by technology could change arbitration as we know it. Both, however, acknowledged that the current arbitration regime does not yet allow for these approaches to be implemented on a wide scale.

Alongside these more ambitious proposals, equally recurring was the idea of using technology to improve existing arbitral practice without an overhaul. In a wider discussion on “Information in Arbitration”, technology was presented as an avenue to address long-standing challenges relating to the quantity and quality of information. Examples included using AI-powered tools to distil large datasets into actionable insights, as well as augmented and virtual reality to present and interact with information in new ways.

At the same time, the use of technology is not without risks. The sea of legal technology tools led one commentator to propose an approach to selecting and using technology in arbitration, to balance technical capabilities with practical needs. Another commentary warned of the unintentional distortion of evidence when using technology. Examples included algorithmic enhancement of videos and images, defective OCR (Optical Character Recognition) affecting document production, and compression techniques that may alter the content of scanned documents.

Artificial Intelligence in Arbitration

Following the release of ChatGPT in 2022, generative AI also entered the stage of arbitration. Early debate focused on the risks posed by AI and on calls to contain its use (see the 2023 Year in Review). One year later, the focus began to shift. The European Union ("EU") enacted the very first AI Act, several arbitral institutions issued guidelines, and attention turned to concrete use cases (see the 2024 Year in Review).

From Experimentation to Normalisation

In 2025, AI remained a dominant topic, but its use appeared increasingly normalised. More than half of respondents to the 2025 International Arbitration Survey expected to use AI “almost always” or at least “often” in the coming years, while only a shrinking minority did not expect to use AI at all.

This normalisation was reflected in two conference reports, which noted increasing pressure from clients to use AI and reduce costs (see here and here). In response, more law firms subscribed to legal AI tools offered by external providers. Over the same period, several legal AI companies reached multi-billion-dollar valuations.

The same enthusiasm can be observed amongst arbitral institutions. The American Arbitration Association (“AAA”) was one of the frontrunners, offering a wide range of AI tools to assist with tasks, such as drafting arbitration clauses, identifying arbitrators and mediators, and generating timelines and case summaries (see here). The AAA also introduced an “AI Arbitrator” for certain document-only construction cases, where an award is drafted by AI and then reviewed and issued by a human arbitrator.

Although acceptance of using AI for legal reasoning remains limited, several contributions explored whether AI might soon expand into new areas, such as automated quantum determination. A strong majority viewed such assistance as acceptable, provided that final decisions remain with human arbitrators.

Whether scepticism toward AI assistance in decision-making will continue remains to be seen. The emergence of platforms promoting either hybrid AI-human models (see e.g. here), as well as fully automated decision-making systems (see e.g. here), suggests that the long-debated “robo-judge” can no longer be dismissed as theoretical or remote. One contribution went further, asking whether AI could eventually redesign the arbitral procedure itself.

Persistent Risks and Regulatory Responses

Despite this wider use, concerns about AI were not dispelled in 2025. Arbitration practitioners continued to identify the risk of undetected errors and bias as principal obstacles to broader adoption. Against this backdrop, one contribution discussed the concept of “Constitutional AI”, which seeks to mitigate these risks by embedding procedural and ethical principles directly into AI systems. This concept is attributed to researchers at Anthropic, which recently announced a new constitution of guiding principles for its language model, Claude.

Lawmakers, too, continued to explore ways to regulate AI. Peru recently enacted an AI law that adopts a risk-based approach similar to that of the EU AI Act. At the same time, 2025 saw growing scepticism toward regulation. In the EU, this manifested itself in extended timelines for the application of certain obligations under the EU AI Act and in efforts to introduce greater flexibility for businesses (see the EU Digital Omnibus). In the US, the administration issued a series of executive orders on AI, including one calling for a review of state-level AI laws to assess their consistency with a “minimally burdensome national framework”.

Institutional Guidance and What May Lie Ahead

Amid growing scepticism towards binding regulation, more institutions issued AI guidelines in 2025, including AAA-ICDR, CIArb, and VIAC. Others – most notably SVAMC and SCC – had already led the way in 2024. While these initiatives respond to a clear demand from arbitration practitioners for guidance, they also attract criticism. One contributor noted that, while all guidelines address whether arbitrators should disclose their use of AI, the approaches vary significantly – from requiring party consent, to mere notification, to no disclosure at all. The contributor questioned whether the guidelines ultimately “arrived too soon,” as no established best practice yet exists for them to reflect.

Looking ahead, it is difficult to predict how arbitration practice will engage with AI in 2026. It is equally difficult to foresee how AI itself will develop over the coming year. Conference discussions in 2025 suggested that law firms may operate with leaner teams, fewer junior lawyers, and greater reliance on AI. Others referred to Amara’s Law, which holds that technological change tends to be overestimated in the short term but underestimated in the long run. Which of these views will prove more accurate remains to be seen.

Technology and the Growing Use of Arbitration for Intellectual Property Disputes

In 2025, arbitration continued to experience growing popularity as a method for resolving intellectual property (“IP”) disputes. While this area of law has been traditionally reserved for national courts, the legal landscape has evolved such that, at the very least, contractual disputes relating to royalties, licensing and non-compete clauses are increasingly considered arbitrable. Contributors from Europe and Asia reported that practitioners in their regions increasingly view arbitration as a viable option for resolving IP-related disputes.

2025’s hot IP topic on the Blog was certainly patent disputes. Previous contributions had already reported on the clear trend across multiple jurisdictions toward the arbitrability of patent disputes (see here and here). That trend was recently reflected in the release of the draft Arbitration Rules for the Patent Mediation and Arbitration Centre (“PMAC”). One post, co-authored by a member of our editorial team, explained how the rules combined state-of-the-art arbitration procedures with patent-specific features, including special procedures for disputes over standard essential patents and their fair, reasonable, and non-discriminatory licensing.

It was not only the promise but also the challenges of the PMAC that made their way into the Blog. The small number of completed cases to date, the small pool of qualified neutrals, and cultural bias towards litigation were identified as challenges in fully embracing the potential of the PMAC as a go-to forum for complex patent disputes. 

Given these developments, we expect to see more on the Blog this year on how IP arbitration fares with the unique challenges thrown its way.

Conclusion

In an age of rapid technological development, the deployment and effect of new technologies undoubtedly raise many thoughts and questions related to dispute resolution. This is clear in the volume of submissions received by the Blog’s technology team in 2025, which shows that arbitration is adapting, step by step, to a changing world. We thank our contributors and readers for helping map this journey and for continuing to shape the future of arbitration.