AI AS A SOURCE OF DISPUTES, NOT JUST EFFICIENCY
CD: What types of disputes are general counsels (GCs) now encountering where artificial intelligence (AI) is the root cause, not just a tool?
Sikellis: It is no longer just about how artificial intelligence (AI) is used – it is about whether the AI system itself created the legal risk. One key area is in the employment setting: algorithmic bias and discrimination claims where AI systems may replicate or amplify existing inequalities. There are also disputes around data usage and intellectual property, including challenges to how models are trained and whether underlying data was used appropriately. There is the possibility of product liability and negligence claims, where AI-driven outputs or decisions allegedly cause harm, raising questions around causation and accountability. And as is the case for any new technology, there is a governance dimension. Companies need to ensure that they are adequately overseeing and controlling for the risks associated with AI.
CD: How are internal policies evolving as AI chat logs, prompts and interactions become discoverable evidence in legal proceedings?
Sikellis: Companies need to create policies and provide training that recognises that AI interactions leave a record in the same manner as email or messaging platforms. Practically speaking, that means companies need to ensure that legal hold, preservation and e-discovery protocols account for AI tools. Companies should consider, where appropriate, limiting or restricting the use of public AI tools, or limiting the use of internal AI tools for sensitive purposes. There should also be training and guidance on prompt ‘hygiene’, setting clear expectations about what should and should not be entered into AI systems.
Jul-Sep 2026 issue
Novartis