eDiscovery has become increasingly complex and costly. What once involved a few servers and email archives has become a vast network of cloud systems, collaboration tools, and AI-driven applications. As the volume and variety of data grow, so do the risks of managing it.

Legal teams are under increasing pressure to balance efficiency with ethics, accuracy, and defensibility in a constantly changing digital environment. Judges are paying closer attention to how emerging technologies affect discovery standards, privilege, and professional responsibility.

The rise of generative AI in eDiscovery has introduced both new efficiencies and new risks, forcing courts and counsel alike to reconsider long-standing assumptions about diligence and review. The challenge now is to define what “reasonable” diligence means when algorithms are part of the review process.

The next phase of discovery will hinge as much on the legal profession’s adaptability as on the capabilities of emerging technology. To succeed, legal teams must earn trust, develop competence, and stay grounded in the ethical standards that guide sound legal work.

Competence and Ethical Responsibility

During a recent TransPerfect Legal panel discussion, judges emphasized that lawyers have an ethical duty to understand how generative AI works. Under Rule 1.1 of the ABA Model Rules of Professional Conduct, competence now extends to technological literacy. Just as lawyers shouldn’t handle a case outside their area of substantive expertise, they shouldn’t use AI tools without understanding how they function and the potential risks they introduce.

Judges emphasized that without a solid understanding of how a system works, lawyers risk failing in core ethical duties like protecting confidentiality, supervising teams, and being candid with the court. As one judge observed, “You can’t ensure confidentiality if you don’t know how the tool works.”

Competence doesn’t mean every attorney must be a tech expert, but it does mean knowing when to ask for help. That might involve working with in-house legal tech teams, outside vendors, or AI specialists who understand model behavior and data security. A key part of modern legal competence is simply knowing the right questions to ask and bringing in the right external resources.

Staying competent also means ongoing learning. With technology rapidly changing the legal landscape, ongoing education is essential. Many bar associations now offer technology-focused CLE programs, which judges encourage lawyers to pursue.

Appropriate Use Policies and Human Oversight

Ethical risks increase when firms lack clear policies around AI use. But when attorneys understand their tools and follow established norms, those risks are easier to manage. A strong use policy should define what AI tools are appropriate to use for business purposes, what use cases they support, and how to validate outputs before relying on them.

The judges consistently stressed that human oversight is essential. Generative AI can accelerate review, but it cannot replace professional judgment. Responsibility still rests with attorneys. It’s not enough to trust confident-looking output. Lawyers must validate the results.

Confidentiality and Client Transparency

Confidentiality is a top concern, especially when working with third-party AI vendors. Judges have emphasized that while technology partners can assist with discovery, lawyers remain ultimately responsible for protecting client data and ensuring compliance with confidentiality agreements.

No formal rule yet requires client consent for AI use, but courts increasingly view transparency as best practice. Clients should know when their data is being used in AI-driven workflows, particularly in sensitive or high-stakes matters.

If a firm uses AI to summarize depositions or prioritize document review, clients should know what safeguards are in place to keep data secure. A brief discussion early in the engagement can prevent difficult conversations later if questions arise regarding how AI was used.

Being transparent from the start helps build trust and can avoid disputes later on, especially as clients become more aware of how AI tools are being used in discovery and review.

From TAR to GenAI: Lessons Learned

The arrival of technology assisted review (TAR) transformed document review, but it also triggered extensive motion practice and skepticism that slowed adoption.

Judges now see parallels between that period and today’s introduction of generative AI, yet they also note key differences:

• Accessibility: Both TAR and GenAI often rely on vendors, though GenAI tools tend to offer more user-friendly interfaces for legal teams.
• Validation: Statistical sampling remains essential in both TAR and GenAI workflows. The output should be verified as a core part of the AI process.
• Transparency: Questions remain about how much disclosure is needed around GenAI tools, workflows, and prompts.

Several judges have cautioned against repeating old mistakes, such as turning new technology into a source of procedural battles (as occurred in the early years of TAR). Discovery, they remind practitioners, is about results. The key question is not “how the sausage was made,” but whether the end product is accurate and complete.

The judges noted that focusing on outcomes over process will streamline discovery. If AI tools show strong validation metrics—like precision and recall—courts may care less about how the results were achieved and more about whether they satisfy discovery obligations.

Judicial Understanding and ESI Protocols

Counsel can spend substantial time negotiating ESI protocols—only to raise disputes later. Early collaboration and clearer documentation can help avoid that wasted effort.

Parties also tend to commit to rigid procedures too early, before testing whether their approach is feasible. The more practical approach is to document agreed-upon principles, follow established ESI protocol best practices, allow for reasonable adjustments, and involve the court only when necessary.

Judges often recommend keeping ESI protocols simple. Instead of locking in every technical detail up front, parties should agree on key elements: data sources, preservation scope, production formats, and communication plans. Technical teams can refine the details as discovery unfolds.

One judge noted that discovery should focus on results rather than process—the key measure being whether productions are complete, defensible, and meet legal standards.

Access to Justice: Promise and Caution

For many individuals and smaller organizations, litigation has become prohibitively expensive. Judges acknowledge that discovery’s cost and complexity can deter valid claims.

They also see potential for generative AI to help close that gap. By automating portions of document review, summarization, and research, these tools could make sophisticated legal processes more affordable—if used responsibly. Judges describe this as one of AI’s most promising aspects.

Still, they caution that the benefits will take time to reach all practitioners. Larger firms and corporate legal teams may see early advantages, while smaller organizations could be left behind as high costs limit access to new tools. Closing that gap will take coordinated efforts in education, policy, and technology affordability.

Legal professionals also play a vital role. If AI frees up time spent on routine tasks, lawyers can redirect their expertise toward pro bono representation, mentorship, and advocacy—actions that help expand access to justice.

Charting the Path Forward

Judges have made it clear: Innovation in eDiscovery won’t be judged solely by speed or savings. It must align with core legal principles like fairness, proportionality, and competence.

Generative AI brings great potential, but also demands caution. Lawyers who understand its limits and communicate openly with clients will be better positioned to meet both ethical expectations and judicial scrutiny. As one judge noted, discovery tools will evolve, but the values of diligence and integrity will always matter.

Ensure your discovery workflows remain accurate, defensible, and aligned with emerging judicial expectations. Connect with TransPerfect Legal’s eDiscovery specialists to learn more.