Maud Piers, Coordinator, GHENT UNIVERSITY "The notion that using AI somehow diminishes professional credibility is dangerous"

LEADERS LEAGUE: How would you describe the current state of international arbitration?

Maud Piers: A great deal is happening, both at the macro and micro level. Geopolitical tensions are reshaping parts of the landscape, but at the same time we see many positive developments that push the arbitration community to reflect, evolve and improve from within – it’s something this community does remarkably well. There is no ivory-tower rigidity; instead, we see strong, self-driven adaptability. Adaptability matters, but never at the expense of principle. This combination of internal responsiveness and principled flexibility positions the field well for the next set of challenges.

Technology has long helped the arbitration community adapt, but AI is now pushing this to new levels. How is the community coping with this shift?

What we see so far is a strong investment in AI literacy: not only understanding what AI can offer, but also having a clear grasp of the risks it introduces. Several initiatives reflect this. The SVAMC and CIArb guidelines, for instance, provide a high-level framework on accountability, risk awareness and proportionate transparency. Importantly, they do so while preserving meaningful discretion for arbitrators and parties. It’s about equipping the community, not constraining it.

Why are these initiatives so critical?

AI literacy although now a legal requirement in Europe is not a box you tick once; it’s a continuous obligation. That said, I don’t believe we need additional legislative frameworks tailored specifically to arbitration. What we do need are frameworks that encourage responsible and transparent use. Clear guidance helps prevent misuse, but just as importantly, it helps prevent secrecy. And secrecy is the part we haven’t talked about enough.

There is still hesitation in professional contexts when it comes to openly acknowledging the use of AI tools. Several recent studies have highlighted this phenomenon, often referred to as shadow AI – showing that lawyers continue to rely on AI informally, outside official workflows. They do this quietly because they fear reputational consequences. A study by HEC Paris pointed at one of the causes of this dynamic: AI-assisted work is initially graded higher, but once it’s known AI had been used, the value is downgraded. This narrative that using AI somehow diminishes professional credibility is dangerous. It drives people into the shadows, where the risks to confidentiality and cybersecurity only increase. And we should absolutely avoid that in arbitration.

What other underexplored risks exist?

Deepfakes pose a fundamentally different kind of threat they introduce active deception, which goes straight to the core of evidentiary trust. We also need to prepare for an increase in disputes about AI within arbitration, and make sure that AI enhances efficiency rather than creating new procedural detours. Another area that deserves far more attention is what some researchers call cognitive debt.

An MIT study, entitled Your Brain on ChatGPT, showed that heavy reliance on AI can weaken neural engagement and reduce memory ownership. When you translate that into an arbitral setting, the questions become much broader than accuracy or productivity.

What happens to our analytical, interpretive or even empathic capacities if we gradually outsource parts of our reasoning to machines?

This ties into a concern we often hear regarding junior practitioners. If AI tools take over large parts of drafting, research or case preparation, we risk hollowing out the traditional training pipeline through which junior lawyers develop the experience that eventually makes them strong senior practitioners.