Annika K. Martin
Sexual-Abuse Litigation · Class Actions · AI Accountability
“AI-generated CSAM is a scourge on society that touches every community and every demographic.
Scale That Serves Choice
A class action is a machine for aggregation, and aggregation cuts both ways in a survivor’s case. Joined to thousands of others, a single voice can move a university or company that ignored it for decades; processed like thousands of others, that same voice can disappear into a claims grid. Annika K. Martin has spent two decades refining an answer to that tension. The cases she leads are engineered so that collective force and individual choice arrive together — settlements in which a survivor decides how much to tell, to whom, and at what depth, without surrendering any share of the leverage the group creates.
Martin is a partner in Lieff Cabraser’s New York office and heads the firm’s Sexual Abuse Survivors Practice Group. She has represented plaintiffs since 2005, and her appointments run through two of the largest university abuse resolutions in American history to a pending case testing whether the newest generative technology can be held to answer for the oldest kind of harm.
A Settlement With Several Doors
The balance she seeks was written directly into the federal litigation over former University of Southern California gynecologist George Tyndall, whose former patients alleged sexual abuse during examinations and a university that failed to act on complaints. As court-appointed co-lead class counsel, Martin helped represent a class of more than 18,000 former patients in a $215 million settlement that received final approval in February 2020.
What distinguished the agreement was its architecture. Rather than push every woman through an identical process, the settlement offered tiered routes to compensation. A class member could claim a payment without recounting intimate details; a woman seeking a higher individualized award could say more through confidential review. Requiring one uniform interview of every claimant would have excluded people who lacked records, feared renewed disclosure, or simply valued privacy more than money. The tiers converted that reality into procedure. Privacy was not a concession wrung from the process — it was written into the remedy, alongside court-supervised reforms in patient protection, complaint handling, oversight, and women’s health services that reached forward to future patients.
One Frame for a Thousand Histories
Martin carried the same design discipline into the litigation concerning former University of Michigan physician Robert Anderson, where a federal judge appointed her interim class counsel. The claims of former patients — more than a thousand of them, spanning decades of examinations — moved through court-supervised mediation toward a resolution announced at approximately $490 million, with roughly $460 million directed to existing claimants and another $30 million reserved for people who might come forward later. The agreement was finalized in September 2022 after claimants completed an allocation process.
A case of that magnitude needed a common frame sturdy enough to carry thousands of claims yet supple enough to honor different medical histories, limitations periods, damages, and choices about disclosure. It had to function as one collective resolution and, claimant by claimant, as a deeply personal reckoning. Martin’s survivor docket extends the pattern beyond flagship universities: she has litigated on behalf of children alleging abuse at Devereux Advanced Behavioral Health facilities and represented former students of the Hotchkiss School.
The Reporter’s Habits
Martin studied journalism at Northwestern’s Medill School, earning her degree in 2001 after a year at Stockholm University, and took her law degree in 2004 from the University of Southern California — the same institution she would later face across the table in the Tyndall case. In law school she served on law review with a focus on women’s legal issues; her early scholarship examined the potential liability of pro-eating-disorder websites, an interest in how systems quietly injure the vulnerable that anticipated her later practice.
The reporter’s habits persist in how she builds a record: each survivor’s story stays specific, defendants’ explanations are tested against their documents and conduct, and the largest case never loses sight of the person at its center. They persist, too, in her command of the machinery of modern litigation. Martin served as a key negotiator on the BP Deepwater Horizon settlements, which together exceeded $11 billion; was court-appointed pro se liaison counsel in the New England Compounding Pharmacy multidistrict litigation, which produced $227.8 million in settlements for victims of the 2012 fungal meningitis outbreak; holds a co-lead appointment in the General Motors air-conditioning defect MDL; and represents residents of Benton Harbor, Michigan, in litigation over lead-contaminated drinking water. She is a longtime leader in The Sedona Conference’s e-discovery working group, a former chair of the American Association for Justice’s class action litigation group, and a member of the American Law Institute.
The Newest Machinery of Harm
In March 2026, Martin stepped into territory no settlement blueprint yet covers. As lead counsel, she filed a federal class action on behalf of three minors alleging that xAI’s Grok image tools generated sexually explicit images and videos of them from ordinary photographs. The complaint asserts federal and California claims and alleges that the company designed, marketed, operated, and profited from a system that could produce such material despite available safeguards. A July 2026 amended complaint added plaintiffs and named Stability AI in connection with the underlying image-generation technology. The case is unfolding now, its allegations untested and contested.
Whatever the outcome, Martin’s work in the xAI suit follows the questions that have shaped her institutional cases: what an organization knew, what it controlled, and what its design choices made possible. In the Grok litigation, those questions focus on the model’s guardrails and access controls. Martin’s complaint asks who created the harmful image and which design choices allowed it to be created and spread so quickly.
That constancy is the signature of her practice. At USC, the question was how a settlement could let 18,000 women participate on their own terms. At Michigan, it was how one frame could hold a thousand distinct histories. In the AI litigation, it is whether the law can reach the architecture of exploitation before the harm becomes ordinary. In Martin’s hands, the class action does what it was meant to do: it lets thousands stand together, and secures reform across an entire institution, without asking any one of them to disappear.