AI and Data Analytics in MDL Litigation

---

Summary

• The increasing size and complexity of MDL proceedings make traditional case management methods increasingly impractical and expensive. AI and data analytics offer powerful tools to manage discovery, analyze claims data, predict litigation outcomes, and facilitate MDL administration and resolution.
• When used responsibly, these technologies can significantly reduce costs, improve efficiency, and preserve judicial resources. Indeed, as MDLs continue to dominate the federal civil docket, the use of AI and data analytics will likely become standard practice in complex litigation.
• The challenge moving forward will be ensuring that these tools are used ethically, transparently, and in a manner that enhances, rather than undermines, the fairness and integrity of the judicial process.

Liubomyr Vorona via Getty Images

---

---

Multidistrict litigation (MDL) now encompasses a commanding 65 percent of the federal civil docket. This explosion in volume has rendered traditional litigation methods functionally obsolete, replacing manageable caseloads with “litigation cities” of thousands of claimants and petabytes of data. To survive this shift, the legal profession is turning to artificial intelligence (AI) as the new standard for case management. From accelerating e-discovery to predicting judicial outcomes, AI and data analytics are no longer just tools in the MDL toolkit—they are the essential infrastructure required to move the modern federal docket toward resolution. This article discusses that shift and how varying AI tools can be leveraged to achieve greater efficiencies.

The scale and complexity of modern MDLs have created significant case management challenges for courts and litigants alike. MDLs have grown from 1 percent to over 65 percent of overall federal court cases, which does not account for state court consolidated or aggregated litigation programs. Nora Freeman Engstrom, Brianne Holland-Stergar & Owen Foulkes, Managing MDLs: A Report from the March 2025 MDL Case Management Convening at Stanford Law School (Rhode Ctr. on the Legal Profession Sept. 2025). Among the most significant challenges are managing massive discovery volumes, ensuring claim integrity, coordinating across hundreds of counsel, and moving cases efficiently toward resolution. See Hamer v. LivaNova Deutschland GmbH, 994 F.3d 172, 178 (3d Cir. 2021) (“A district court, administering a multidistrict case, faces unique challenges not present when administrating cases on a routine docket.”). This places significant strain on judicial resources and increases litigation costs for all parties, but most notably the producing parties. See, e.g., ** In re Guidant Corp. Implantable Defibrillators Prods. Liab. Litig., 496 F.3d 863, 868 (8th Cir. 2007) (“Given the time pressure on a defendant that must investigate the claims of nearly 1,400 plaintiffs, we consider the danger of prejudice substantial.”).

Attorneys and courts are leveraging AI and other data analytics tools to manage large MDL dockets more efficiently and strategically. Melissa Whitney, Bellwether Trials in MDL Proceedings 31 (Fed. Jud. Ctr. 2019) (recommending that courts “encourage the parties to adopt electronic data-collection tools to better track and understand the cases that have been filed and to help identify major variables on which individual case outcomes may turn”). Indeed, “it is indisputable that generative AI may automate many time-consuming tasks in the eDiscovery process, such as significantly accelerating document review, quickly summarizing documents, streamlining privilege review and logging, identifying named entities like key people and organizations, extracting important topics, and automating writing tasks[.]” Jeffries v. Harcros Chems. Inc., 2026 U.S. Dist. LEXIS 63182, at *14 (D. Kan. Mar. 25, 2026).

E-Discovery and Document Review

One of the most significant applications of AI in MDL practice is in electronic discovery. MDLs involve millions of discovery documents, including emails, scanned files, corporate records, medical records, and expert materials. AI-powered technology-assisted review, commonly referred to as predictive coding, uses machine learning algorithms to identify relevant documents and prioritize them for review, significantly reducing the time and cost associated with manual document review. Maura Grossman & Gordon Cormack, “ Technology-Assisted Review in E-Discovery Can Be More Effective and More Efficient Than Exhaustive Manual Review,” 17 Rich. J.L. & Tech. 11 (2011). Courts have increasingly accepted predictive coding as a proportional and efficient method of discovery review, recognizing that it can be quicker and more accurate than traditional manual review in large-scale litigation, and it remains a challenging element of protocol negotiations involving electronically stored information (ESI).

Discovery obligations in MDLs are centralized and involve common issues across thousands of cases, and certain uses of predictive coding allow parties to efficiently identify key custodians, relevant document sets, and privilege materials, thereby streamlining discovery and reducing litigation costs. Bogdana Stjepanović, Leveraging Artificial Intelligence in eDiscovery: Enhancing Efficiency, Accuracy, and Ethical Considerations 179–94 (Inst. Compar. L. 2024).  In addition, cybersecurity and data privacy challenges need to be addressed in any protective orders, ESI protocols, and end-of-life-cycle decommissioning protocols.

AI is not perfect, just like any review method, and confidence calculations need to assess whether relevant or nonresponsive documents are captured. Privilege determinations in particular still require careful attorney review and quality-control measures. Parties should address the use of AI in discovery protocols at the outset of litigation. Indeed, parties often confer on custodians, search terms, and production formats—there is no reason those discussions cannot also include the scope and acceptable use of AI in document review.  Reaching agreement on these issues beforehand and documenting them in a discovery protocol may help avoid future disputes regarding the adequacy of document review, privilege screening, and production methodology.

---

---

---

---

Data Analytics and Generative AI Engagement for Enhanced Case Management and Strategy

Beyond discovery, data analytics plays an increasingly important role in MDL case management. MDLs generate enormous amounts of structured and unstructured data, including plaintiff fact sheets, medical records, injury data, and docket filings. Data analytics platforms allow parties and courts to analyze trends in claims, injuries, and demographics, which can inform decisions regarding discovery pools, bellwether selection, and settlement strategy. See Angela Browning, “ Using Data Analytics in the Management of MDLs,” Litig. Mgmt., Inc., Blog (2020) (describing data analytics as “a largely untapped, yet powerful tool in the management of product liability MDLs” that “can help parties and judges make better decisions in how MDLs are managed going forward”).

Early census and plaintiff fact sheet data, for example, can be analyzed to identify claim validity, injury patterns, and potential case valuation ranges. This data-driven approach allows courts and parties to identify weak claims early, streamline discovery, and develop more efficient resolution frameworks. Data analytics can reduce the number of non-meritorious claims that remain in the MDL and assist courts in managing large dockets more effectively. Analysis of data points from early case-dispositive motions, settlement agreements, and bellwether verdicts can promote a more rapid global resolution.

Generative AI tools offer practical support for day‑to‑day MDL case management. Generative AI platforms can produce outlines of complex issues, prepare summaries of lengthy court filings or deposition transcripts, and generate timelines based on docket entries or key evidentiary materials. These tools may assist counsel in identifying themes across large document sets, synthesizing factual patterns in discovery, and organizing information for case management conferences or bellwether briefings. Although such outputs require attorney oversight to ensure accuracy and strategic alignment, generative AI’s ability to accelerate these foundational tasks may help case teams focus on higher‑level litigation strategy.

Predictive Analytics and Litigation Outcomes

AI tools are also leveraged to predict litigation outcomes and assist in strategic decision-making. By analyzing historical docket data, prior judicial rulings, and case outcomes, machine learning models can identify patterns that help attorneys evaluate case value, settlement ranges, and trial risk. See Ashley Hallene & Jeffrey M. Allen, “ Using AI for Predictive Analytics in Litigation,” Voice of Experience, Oct. 16, 2024 (noting that “[p]redictive analytics tools can then analyze historical data from similar cases to predict the likely outcome of the new case,” including “the probability of winning or losing, potential settlement amounts, and the duration of the case”). These predictive analytics tools can assist in selecting bellwether cases, evaluating dispositive motion strategy, and determining whether early settlement may be advantageous. Id.; James Sullivan & Abigail Cahn-Gambino, “ AI Platforms Can Help Product Liability Litigation Move Forward,” Bloomberg L., Jan. 15, 2026 (noting that “[l]itigants can use [AI] tools to support a data-driven MDL that helps shape important decisions on which cases to work up through discovery pools, selection of potential bellwether trials, whether to implement Lone Pine or similar orders, and other resolution opportunities”).

But the clever litigator does not merely consider these analyses as a binary result—i.e., ** to settle or not to settle. Litigation analytics platforms can further predict key benchmarks in the litigation that help determine the best time to settle. Pre/Dicta, “ AI-Powered Legal Case Outcome Prediction: Transforming Legal Practice,” Feb. 28, 2025. For example, the platforms can provide insights into judicial tendencies, motion success rates, and chronologies for future key events in the MDL. If the predictive model suggests a defense attorney’s pending motion to dismiss claims from a master complaint will likely fail but that a future motion for summary judgment on certain products liability claims is more likely to succeed, simultaneously narrowing all plaintiffs’ cases across the MDL, it may make more sense to wait until that period of the litigation to engage in settlement discussions. Attorneys will often consider many factors such as this in deciding when to negotiate or mediate claims, but predictive analytics further informs those choices and helps justify them to case teams and clients.

Outside of settlement, the same insights enable more informed strategic decisions in complex MDL proceedings. Where analytics predict a motion to dismiss products liability claims is more likely to be granted than a motion to dismiss parallel negligence claims, which has a low chance of success, the moving party will likely want to devote more of the page limit to emphasizing products-related points. In an instance where the model indicates the judge is inclined, based on prior rulings, to exclude certain types of medical records as substantially more prejudicial than probative, it could be beneficial to spend more time at a plaintiff’s deposition building out the relevance of the key records you will seek to use at trial to strengthen the basis of their admissibility. And where a platform estimates that an order on a motion to dismiss certain bellwether complaints will be received shortly before a case management conference, it would be wise to prepare arguments for alternative bellwether selections prior to that conference. While these tools do not replace attorney judgment, they provide valuable data-driven insights that can improve decision-making in large-scale litigation.

Ethical and Practical Considerations

Despite the benefits of AI and data analytics, use of these tools in MDLs continues to raise ethical and practical concerns. Attorneys must ensure compliance with professional responsibility obligations, including competence, confidentiality, and supervision of technology-assisted work product, as set forth in Model Rules of Professional Conduct 1.1, 1.3, and 1.6. Overreliance on AI tools may also lead to errors, including inaccurate document classifications or flawed predictive models, particularly if training data are incomplete or biased. Therefore, courts and practitioners must ensure that AI tools are used transparently and with appropriate human oversight. The emerging consensus remains that AI should be used to augment, rather than replace, attorney judgment and judicial decision-making. In addition, counsel need to work carefully with their clients to abide by internal company guidelines, protocols, approved uses, and any land mines associated with data management, commingled data, or what to do when a litigation ends.

The increasing size and complexity of MDL proceedings make traditional case management methods increasingly impractical and expensive. AI and data analytics offer powerful tools to manage discovery, analyze claims data, predict litigation outcomes, and facilitate MDL administration and resolution. When used responsibly, these technologies can significantly reduce costs, improve efficiency, and preserve judicial resources. Indeed, as MDLs continue to dominate the federal civil docket, the use of AI and data analytics will likely become standard practice in complex litigation. The challenge moving forward will be ensuring that these tools are used ethically, transparently, and in a manner that enhances, rather than undermines, the fairness and integrity of the judicial process.

---

Author

Farwa Tahir

...

View Bio →

Thomas Andrew Zelante Jr.

...

View Bio →

Kaitlyn Stone

Barnes & Thornburg LLP

...

View Bio →

Michael Charles Zogby

Barnes & Thornburg LLP

...

View Bio →

---

Author

Farwa Tahir

Thomas Andrew Zelante Jr.

Kaitlyn Stone

Barnes & Thornburg LLP

Michael Charles Zogby

Barnes & Thornburg LLP

---

Committees

This content was produced by:

• Mass Torts Litigation Committee

Related Content

Appellate Practice | Pretrial Practice | The Practice of Law

Failure to Verify: Why the Basics Matter Most

Apr 14, 2026

The Practice of Law

Generative AI and Privilege: Practical Lessons from Two Early Decisions and What Comes Next

Mar 19, 2026

The Practice of Law

GenAI and the Ethical Lawyer: Same Standards, Different Dilemmas Part II—How Courts Are Enforcing the Duty of Candor

Mar 10, 2026

The Practice of Law

Use of Generative AI in the Law: Lessons from Two Federal Cases

Mar 10, 2026

Intellectual Property & Technology Litigation | The Practice of Law

Federal Court Rejects Privilege Assertion for AI-Generated Research

Mar 05, 2026

Civil Litigation | The Practice of Law

AI Communications Are Not Privileged: What United States v. Heppner Means for Corporate Clients

Feb 27, 2026