After apparently using GenAI for privilege review, “Meta downgraded its privilege designations on 56,747 of the 60,462 documents it had reviewed, or about 94%, according to the plaintiffs’ motion on Friday. The plaintiffs say the results show Meta deliberately used attorney-client privilege to keep highly relevant evidence out of discovery.” C. Clouogh, Meta’s ‘Privilege Abuse’ Warrants Sanctions, MDL Judge Told – Law360 (Nov. 16, 2026).
“The scale, persistence, and direction of these reversals support the conclusion that Meta applied the privilege far too broadly and did so deliberately, rather than making isolated good-faith mistakes,” the plaintiffs say.
Id. The article reports that Meta has reserved a clawback right. It adds that plaintiffs “ ask the court to let them use the downgraded documents with any witness or expert and to bar Meta from raising objections based on authenticity, foundation, personal knowledge or hearsay, while preserving objections under Federal Rule of Evidence 403. They also ask for a jury instruction allowing jurors to infer that Meta withheld the documents because it believed they were unfavorable to the company.”
According to Plaintiffs, “Meta cites its use of generative artificial intelligence in arguing it should retain the right to claw back documents. The plaintiffs ask the judge to require Meta to seek court permission and show good cause before clawing back any document produced through the review.”
Plaintiffs allege that Meta hid the risks of teen social media use. The MDL is In re: Social Media Adolescent Addiction/Personal Injury Products Liability Litigation, case number 4:22-md-03047, in the U.S. District Court for the Northern District of California.
This post is based on the Law360 article.