As AI systems become more capable, autonomous, and embedded in business operations, the potential for AI-related disputes continues to grow. At the center of many of these matters will be technical questions about how AI systems were developed, tested, deployed, controlled, and ultimately behaved. Because AI encompasses a range of technologies and disciplines, the expertise needed will depend on the systems and technical questions at issue. Understanding those questions can be critical to identifying the right evidence, evaluating competing claims, and explaining complex technology to decision-makers.
Here are ten considerations for counsel to bear in mind when evaluating the AI expertise needed in impending litigation.
- Understand the AI at issue. Not all AI systems work the same way, and generative AI and large language models represent only part of the broader AI landscape. An expert can help identify the type of AI system, its architecture, underlying data, intended function, and limitations, providing the technical foundation for evaluating the issues in dispute.
- Reconstruct what the system actually did. AI-related disputes may turn on understanding how a system produced an output or took an action. An expert can analyze prompts, outputs, logs, model versions, system configurations, and other available technical evidence to help reconstruct what occurred.
- Separate model behavior from human and system factors. An unexpected outcome does not necessarily mean the AI itself failed. Expert analysis can help distinguish model behavior from data quality, system design, integration issues, and how people interacted with or relied on the technology.
- Examine how the AI was tested and evaluated. Claims that a system was tested, validated, or monitored can mean very different things. An expert can assess the methodology, benchmarks, performance measures, safeguards, and limitations underlying those claims.
- Understand the data behind the system. Training, testing, and operational data can influence how an AI system performs. An expert can evaluate the quality and limitations of relevant data and its potential role in the behavior at issue.
- Consider the role of human-AI interaction. How people interpret, rely on, override, or respond to AI-generated information can be important to understanding an outcome. Experts in human-AI interaction can evaluate the relationship between system behavior and user behavior when it is relevant to the dispute.
- Look beyond the output when AI-generated content is disputed. Text, images, audio, and other AI-generated content can raise different technical questions. Experts with the appropriate specialization can analyze how content was generated, modified, or processed and the technical evidence relevant to evaluating it.
- Evaluate the controls and safeguards surrounding the AI. As AI systems become more capable, questions may arise about how their behavior was constrained, monitored, and governed. An expert can assess the technical controls, testing, risk management processes, and safeguards in place and explain their capabilities and limitations.
- Understand the added complexity of autonomous and agentic AI. Systems that can use tools, access other systems, make decisions, or execute multistep tasks introduce additional complexity. Experts can help reconstruct their actions and evaluate the system architecture, interactions, controls, and safeguards governing their behavior.
- Identify the right technical evidence early. The information and expertise needed to evaluate an AI-related dispute may not be obvious at the outset. Early expert involvement can help counsel identify relevant technical records, recognize gaps in available information, and frame the technical questions that may shape discovery. Depending on the systems and issues involved, a matter may also require expertise across AI disciplines rather than a single broadly credentialed expert.
As AI-related disputes become more common and complex, understanding the technology at issue will be increasingly important to building an effective litigation strategy. The right expertise and AI expert team can help counsel identify critical evidence, evaluate technical claims, and translate complex AI systems into issues decision-makers can understand.
