How Courts Are Addressing AI-Generated Content in Discovery Disputes
Generative AI has been transforming the legal industry, assisting with everything from case law summarization to document review and legal research. But as its adoption accelerates, new legal questions are emerging:
- Should AI-generated content—including prompts and outputs—be subject to discovery?
- Can opposing counsel request AI-generated legal research as part of a case?
- Do AI developers have to disclose their training data in litigation?
In 2024, courts began grappling with these complex questions, and their rulings are already shaping the future of AI in litigation.
How AI Is Changing eDiscovery
Litigators are increasingly using AI to:
- Summarize depositions and case law.
- Draft legal arguments.
- Assist with document review and privilege determinations.
But AI also raises serious risks, particularly when it comes to transparency and discoverability.
Key Case Law Developments on AI and Discovery
1. AI-Generated Content May Be Discoverable
For the first time, courts have begun considering whether AI-generated legal research, deposition summaries, and internal firm memos should be subject to discovery.
Since January 2024, fifteen decisions in eDiscovery Assistant were tagged with Generative AI, reflecting a growing trend of courts addressing these issues. While no firm precedent has been established yet, these cases suggest that AI-generated content will not automatically be shielded from discovery.
Takeaway: If a law firm uses AI to summarize case law or draft internal legal arguments, opposing counsel may seek discovery of those outputs—raising major privilege concerns.
2. AI Training Data Is a New eDiscovery Battleground
Recent lawsuits against OpenAI, Meta, and Microsoft have introduced a new eDiscovery challenge: Can companies be forced to disclose the training data used to build their AI models?
These disputes raise critical questions about trade secrets, privacy, and copyright law. Some courts have allowed discovery into AI training datasets, arguing that if AI outputs are at issue in a case, the training data behind them could be relevant.
Takeaway: AI training data may soon become a standard eDiscovery request in IP and privacy litigation. Companies using AI should anticipate potential legal battles over data sources.
What This Means for Litigators
- Evaluate AI Use in Your Cases – If opposing counsel relies on AI-generated content, should it be produced in discovery?
- Address AI in ESI Protocols – Clarify whether AI-generated summaries, prompts, or legal research must be retained.
- Consider Ethical Implications – Courts are still navigating AI’s role in privilege, accuracy, and bias—issues that could impact your litigation strategy.
The Future of AI in eDiscovery
The rise of AI in litigation is inevitable, but the rules governing its use are still developing. As courts refine their approach, litigators must be proactive in understanding how AI-generated content fits into their discovery obligations.
AI is not just a tool for efficiency—it’s becoming part of the evidence landscape. Staying ahead of the legal and ethical implications will be critical for any litigator navigating this new frontier.
