Courts confront privilege, AI records, and reliable legal work

New disputes and proposals address privilege reviews, ESI scope, AI-use records, automated authorship claims, and fabricated judicial authority.

Edition
eDiscovery Decoder Newsletter

Sent November 18, 2026

Lead story

Meta faces sanctions motion over privilege designations

Plaintiffs in the social-media addiction litigation asked a federal judge to sanction Meta for allegedly designating thousands of documents as privileged without adequate support.

What happened
Plaintiffs in the social-media addiction litigation asked a federal judge to sanction Meta for allegedly designating thousands of documents as privileged without adequate support.
Why it matters
Large privilege reviews need repeatable legal criteria, documented quality control, and escalation for uncertain calls. When a high percentage of challenged designations changes, courts may scrutinize the review process, the adequacy of the log, and whether the designations delayed access to evidence needed for trial.
Watch
The sanctions request is pending. The percentages and characterization of Meta's conduct are the plaintiffs' allegations, and the court has not found misconduct.

SourcesLaw.com

Briefing item

Court narrows initial ESI exchange in Apple-OpenAI dispute

A Northern District of California judge directed Apple and OpenAI to continue negotiating custodian limits but declined to require project codenames in their initial ESI exchange.

What happened
A Northern District of California judge directed Apple and OpenAI to continue negotiating custodian limits but declined to require project codenames in their initial ESI exchange.
Why it matters
Discovery scope should follow the purpose of each phase rather than automatically carrying early custodian choices into merits discovery. Staged interrogatories can also test the need for sensitive identifiers before parties undertake a broader collection that risks exposing unrelated confidential projects.
Watch
The order resolves discovery protocol questions, not the merits of Apple's trade-secret claims. The court left open renewed codename discovery if Apple develops additional supporting facts.

Sourceslegal.economictimes.indiatimes.com

Briefing item

Maryland discipline case relies on a lawyer's ChatGPT logs

Maryland's Attorney Grievance Commission filed a disciplinary petition alleging that a lawyer used nonexistent authorities in legal work and used ChatGPT to help answer bar counsel's inquiries.

What happened
Maryland's Attorney Grievance Commission filed a disciplinary petition alleging that a lawyer used nonexistent authorities in legal work and used ChatGPT to help answer bar counsel's inquiries.
Why it matters
Prompts and chat histories can become evidence of authorship, knowledge, and representations made during an investigation. Legal teams should understand platform retention and export settings, preserve relevant AI-use records when a dispute arises, and independently verify every authority before filing or responding to regulators.
Watch
The allegations have not been adjudicated, and the lawyer had not filed a formal answer when the report was published. The disciplinary petition also includes client-service and fee allegations beyond AI use.

Sourcesthedailyrecord.com

Briefing item

Federal prosecutors cite AI detector results in court filings

404 Media reports that Justice Department lawyers used the Pangram AI-writing detector to argue that opposing filings were generated with AI.

What happened
404 Media reports that Justice Department lawyers used the Pangram AI-writing detector to argue that opposing filings were generated with AI.
Why it matters
When AI authorship is disputed, counsel may need to preserve drafts, research history, prompts, outputs, metadata, and reviewer approvals. Detector scores should be evaluated alongside primary evidence because an automated classification does not by itself establish who wrote a document or whether its legal content is reliable.
Watch
The detector result is an advocacy claim in contested litigation, not a judicial finding of AI authorship. 404 Media's report does not establish an independently validated error rate for Pangram in this use.

Sources404media.codocumentcloud.org

Briefing item

New York bill targets fabricated authority in judicial rulings

Proposed New York legislation would define a judge's knowing or reckless inclusion of fabricated legal authority in a ruling as judicial misconduct.

What happened
Proposed New York legislation would define a judge's knowing or reckless inclusion of fabricated legal authority in a ruling as judicial misconduct.
Why it matters
Courts need verification controls for AI-assisted research and drafting that reach judges and staff as well as advocates. Source-level citation checks, retained research trails, and defined review responsibility can help prevent fabricated authority from entering an order and preserve a record if an error is challenged.
Watch
This is proposed legislation, not current law, and its text may change or fail to pass. The Reuters page was inaccessible during verification; a complete Reuters syndication was reviewed through SRN News.

Sourcesreuters.comsrnnews.com

Educational only. Not legal advice. Review sources and matter-specific obligations before relying on any workflow decision.