Innovation Leaders from A&O Shearman, White & Case, Signature Litigation, and Legora Discuss Governance, Adoption, and Accountability in Legal AI Deployment
THE AUTHOR:
Tiffany Lam, Strategic Communications and AI Ethics Officer at Jus Mundi
During Paris Arbitration Week 2026, Jus Mundi convened innovation leaders at Le Meurice for its AI for Justice Symposium to address a critical question: How is ethical AI operationalised in legal practice?
The panel moderated by the CEO and Co-Founder of Jus Mundi, Jean-Rémi de Maistre, brought together Lucas De Ferrari (Partner, White & Case), Natalia Chumak (Partner, Signature Litigation), Eleissa Karaj (Paris Innovation Manager, A&O Shearman), and Filip Nordlund (Legal Engineering Manager, Legora) to discuss what happens when AI strategy becomes AI reality, focusing on the adoption of legal tech tools, training, governance, and accountability.
At the symposium, Jus Mundi and Legora also announced the first agent-to-agent integration in arbitration, enabling practitioners to query Jus AI directly within Legora’s workspace. The conversations that followed illustrated why such innovations matter and what obstacles they must overcome.
From Strategy to Reality: Governance Frameworks
The panel discussed distinct approaches to AI governance, each revealing what successful implementation requires.
A&O Shearman deployed Harvey firm-wide to 5,000 lawyers across 43 offices in February 2023, following a November 2022 pilot. Eleissa Karaj outlined the firm’s governance approach with a dedicated global AI team, a committee evaluating new tools in the market, and a centralised intranet housing all AI policies and client restrictions, among other initiatives. The key success factors were senior advocates, comprehensive practice-specific training, and informal champions from each practice group.
White & Case announced a global Legora rollout across 43 offices in December 2025, which was one of the largest enterprise legal AI deployments to date, with over 85% license adoption. Lucas De Ferrari emphasised that successful deployment is “mindset rather than just technology”. The impact has been concrete: significantly improved first draft quality, reduced review loops, and faster review processes.
At Signature Litigation, a disputes-only boutique firm, Natalia Chumak noted that AI has the potential to improve the delivery of legal work throughout the whole cycle. She highlighted several concrete examples: pre-claim assessments where AI can reduce upfront costs, urgent matters requiring rapid case understanding, legal research where AI serves as a sounding board for generating new perspectives, and extraction of specific information from dense documents such as expert reports.
Chumak also outlined key considerations in Signature’s evaluation of AI vendors, including transparency around what each tool is designed for and access to vendor analytics to monitor usage patterns; explainability regarding how tools reach answers and their limitations; and consistency in assessing whether similar prompts produce comparable responses.
From a vendor’s perspective, Filip Nordlund described Legora’s legal engineering role as bridging product and practice: showing clients how to use the platform in their specific practice areas, learning from user patterns to improve the product, conducting thousands of evaluations to ensure consistency, and implementing technical improvements based on practitioner feedback. This approach informed the first agent-to-agent integration in arbitration with Jus Mundi – enabling specialised arbitration intelligence to be accessed without workflow disruption, with citations displayed directly alongside answers and Jus Mundi’s data presented uncorrupted within Legora’s interface.
When asked to identify the biggest obstacles to AI deployment in law firms in one word, “inertia” came up. The agent-to-agent integration between Jus Mundi and Legora aims to reduce this friction by eliminating the need to get accustomed to entirely new platforms, allowing lawyers to focus on developing critical AI judgment rather than mastering new interfaces.
The Human Element: Training and Adoption
Beyond governance structures, the panel explored how firms are navigating the human dimensions of AI adoption through training.
Training junior lawyers sparked particularly rich discussion. Karaj shared that trainee lawyers demonstrated strong enthusiasm for experimenting with AI capabilities, with some citing access to advanced AI tools as a deciding factor in choosing which law firms to apply to. A&O Shearman balances this enthusiasm with training frameworks and considerations. The current approach includes deployment-focused training on capabilities and limitations, training senior lawyers to supervise trainee lawyers’ AI use appropriately, and practice-specific approaches tailored to different teams’ needs.
Signature Litigation’s junior lawyer training includes AI literacy from the perspective of a practitioner, tool-specific training emphasising stress testing and critical judgment, as well as AI clinic sessions where providers and lawyers collaborate to solve specific problems.
For AI providers, Nordlund noted that effective training support requires use case-specific approaches, transparency about technical limitations such as document processing constraints, and safeguards around confidentiality, particularly preventing sensitive information from leaving secure platforms through features like web search.
When asked for one word describing successful AI adopters, the panelists responded: agile, open-minded, experimentation, survival. These responses underscored that the challenge of AI adoption is not purely technical but is also cultural.
Accountability and the Path Forward
The discussion’s final section confronted practical realities about vendor dependency, algorithmic decision-making, and institutional accountability. The panel acknowledged that these issues present both opportunities and risks that require ongoing vigilance and community dialogue.
The symposium revealed a clear pattern: successful AI deployment in legal practice depends less on technological sophistication and more on transparent governance, robust training, and accountable partnerships. The obstacles identified are fundamentally human challenges, not merely technical ones.
Events like this symposium address knowledge gaps and share good practices through open dialogue across the arbitration community. As these technologies continue to evolve, the community’s commitment to ethical deployment will determine whether AI serves justice or merely efficiency. At Jus Mundi, our commitment as a Société à Mission is to ensure these tools democratise access to legal intelligence while upholding the highest standards.
The playbook for ethical AI in legal practice is being written in real time. These conversations are not conclusions but continuations; those writing it together – law firms, technology providers, institutions, and the broader legal community – will shape whether AI advances global justice or creates new barriers to it.
About Jus Mundi
Founded in 2019 and recognized as a mission-led company, Jus Mundi is a pioneer in the legal technology industry dedicated to powering global justice through artificial intelligence. Headquartered in Paris, with additional offices in New York, London, and Singapore. Jus Mundi serves over 150,000 users from law firms, multinational corporations, governmental bodies, and academic institutions in more than 80 countries. Through its proprietary AI technology, Jus Mundi provides global legal intelligence, data-driven arbitration professional selection, and business development services.
*The views and opinions expressed by authors are theirs and do not necessarily reflect those of their organizations, employers, or Daily Jus, Jus Mundi, or Jus Connect.
