BELGIUM Trends and Developments Contributed by: Nathalie Colin and Florence Frühling, Freshfields LLP
comes and assist with strategic planning. It sup - ports preliminary assessments, enables the com - parison of legal arguments and helps anticipate legal trends. • “Generative AI” assists in drafting procedural docu - ments, legal submissions and even arbitral awards. The tool also excels at summarising large volumes of information, helping users extract key points from transcripts, witness statements or lengthy legal submissions. • “Operational AI” is designed to streamline case management. The technology automates adminis - trative tasks such as scheduling, deadline tracking, document categorisation and workflow manage - ment. This reduces the time that legal practitioners spend on routine processes and helps lower costs for clients. • “Interactive AI” includes real-time tools for tran - scription, translation, chat-based legal research and virtual hearing support. The tool is especially valuable in multilingual proceedings, ensuring clear communication and accurate documentation. Together, these tools are becoming increasingly embedded in the international arbitration workflow. While adoption rates vary by jurisdiction and institu - tion, the global trend points toward a growing reliance on AI to improve efficiency, consistency and access to justice in international arbitration. Looking ahead, AI is expected to play a more strategic role in early-stage case analysis, allowing parties to analyse the strengths and weaknesses of their claims at an early stage. This can lead to quicker settlements or more focused proceedings. Moreover, advocacy styles may evolve to become more data-driven, with arguments tailored to AI-identified persuasive pat - terns. AI also allows for more strategic arbitrator selection by analysing arbitrators’ previous awards. Challenges in the use of AI in international arbitration Despite its clear advantages, the integration of AI into arbitration raises important legal, ethical and practical concerns.
Confidentiality Unlike court proceedings, arbitration is typically pri - vate and does not produce public precedents. Confi - dentiality obligations and data protection laws restrict access to case materials, making it difficult to train or improve AI systems. Yet, AI tools rely on large datasets to function effectively. This creates a tension between protecting sensitive information and leveraging data- driven technologies. Moreover, AI systems may gener - ate so-called “hallucinations” (ie, where systems gen - erate false or non-existent legal sources). Arbitration practitioners must therefore remain vigilant, verifying AI-generated content and carefully controlling what data is shared with such tools. Transparency Many AI models (especially large language models) operate as “black boxes”, producing results without clear explanations of their reasoning. Such lack of explainability can be problematic in arbitration, where awards must be reasoned and subject to scrutiny. Arbitrators must be able to understand and justify the basis of any AI-generated input they rely upon. Ethical considerations Should arbitrators or the parties’ counsels disclose their use of AI when drafting awards or preparing submissions? Transparency is crucial, particularly when legal reasoning is partially machine-assisted. Additionally, the principle of non-delegation must be upheld: AI may support administrative tasks but must never replace human judgement in assessing evidence, applying legal principles or drafting final awards. Bias and misinformation If AI systems are trained on flawed or unrepresenta - tive datasets, they may perpetuate existing biases, undermining fairness and procedural equality. Arbitrator selection Although parties are free to choose arbitrators, it remains unclear whether an AI system (or its develop - er) can meet essential criteria such as independence, impartiality and neutrality. The opaque nature of the underlying data makes it difficult to assess whether these standards are met, potentially jeopardising the enforceability of awards.
63 CHAMBERS.COM
Powered by FlippingBook