George Hannah is a solicitor apprentice in London, entering the third year of his apprenticeship, alongside writing on LinkedIn and producing a newsletter and podcast focused on how AI is changing the legal profession. Liam Fitter is a Product Compliance Analyst at Dataships, working across GDPR, e-privacy, the EU AI Act and international marketing regulation. Their different perspectives, Hannah tracking developments across legal AI and Fitter applying technology and regulation within a growing business, shaped a practical discussion about what the latest wave of AI developments means for legal professionals.

Overview

The discussion began with AI transparency and the EU AI Act, focusing on when organisations must disclose AI interactions and identify AI-generated content. Fitter highlighted the challenge of determining the relevant regulatory category and assessing risk, particularly for smaller organisations without in-house legal teams. Hannah also noted that disclosure may affect how consumers trust and engage with AI-generated material.

The speakers then explored the expansion and integration of legal AI platforms. Hannah discussed growing interest from major technology companies, while Fitter highlighted connectors linking AI tools with systems such as iManage, NetDocuments, Everlaw and Thomson Reuters, alongside platforms including Harvey and Legora. These integrations can make AI more useful, but also increase the need to control access to sensitive information. The discussion suggested that many vendors are competing to become the central workspace for legal research, drafting and collaboration.

A final theme was the economics, reliability and governance of AI adoption. The speakers discussed unpredictable token consumption, usage-based pricing and the difficulty of forecasting AI costs. They also addressed AI sycophancy, where models may reinforce users’ assumptions, especially when prompts lack important legal context such as jurisdiction. Effective adoption therefore requires informed users, strong governance, careful verification and continued human judgement.

Key Takeaways

• AI transparency is becoming a compliance requirement. Organisations must know when AI interactions and content need to be disclosed.
• More integrations require stronger governance. Organisations must control what data AI systems can access.
• Legal AI providers are competing to become central platforms for research, drafting and workflow management.
• Usage-based pricing makes AI costs harder to predict and manage.
• Convincing AI output may still be unreliable, especially when prompts lack context.
• Human judgement, training and oversight remain essential.