
Share
A hospitality startup's lawsuit against a rival reveals a new wrinkle in trade secrets litigation: chat logs allegedly show an employee asking AI how to steal data without getting caught, and getting warned off.
A lawsuit filed in New York state court this week offers a preview of how AI chat logs may become standard evidence in corporate espionage cases. Hospitality and events platform Dorsia accuses rival startup Rumor Group of poaching an engineer who allegedly used his employer-paid Claude account, built by Anthropic, to ask the chatbot how to steal company secrets and cover his tracks.
The engineer, Chris Mathew, allegedly used Dorsia's account to query Claude on how he could leave for Rumor "without legal exposure," what materials he could take with him, and how to avoid detection. According to the complaint, the chatbot pushed back. "Don't bring anything from [Dorsia] to Rumor. No code, no docs, no playbooks, no contact lists," Claude reportedly told Mathew, in language the lawsuit quotes directly.
Dorsia's filing leans hard on that exchange. "Even AI knew this was wrong; Mathew and Rumor did not care," the complaint states. It is an unusual framing for a trade secrets case: the plaintiff is using the defendant's own AI assistant as a character witness against him.
The complaint also alleges Mathew used the Claude account to sketch out competing event concepts while still employed at Dorsia, with prompts like "events by dorsia but better for nightclubs come up with a sexy luxury name." Dorsia says this shows premeditation, not a one-off lapse in judgment.
Rumor's attorney, Maurice Pessah of Pessah Law Group, rejected the allegations in a statement, saying the two companies "operate fundamentally different businesses, and no amount of spin is going to change that." Pessah countered that Dorsia had made "lowball offers to acquire Rumor under threat of litigation" and called the lawsuit "devoid of merit." Dorsia did not respond to those counterclaims by publication time. Mathew could not be reached for comment.
Dorsia is seeking unspecified monetary damages along with an injunction barring Rumor from using any of its confidential information. Steven Feldman of Latham & Watkins represents Dorsia; the case is Dorsia v. Rumor Group Corp, filed in the Supreme Court of the State of New York, County of New York.
Enterprise AI adoption has outpaced the legal frameworks meant to govern it. Companies now routinely provision employees with paid AI assistants for coding, research, and planning. Those tools retain conversation logs. That creates a discoverable record of intent that did not exist in prior generations of trade secrets disputes, where proving what an employee was thinking before a departure often came down to circumstantial evidence or forensic analysis of file transfers.

This case flips the usual script. Normally, companies worry that employees will feed proprietary data into an AI tool and leak it externally, the classic insider threat scenario security teams have spent the past two years building controls against. Here, the risk ran in the opposite direction. The AI tool itself generated a paper trail that the employer is now using as evidence against a former staffer who allegedly tried to enlist it as a co-conspirator.
The implication for corporate counsel is straightforward: AI chat logs are discoverable, and they may cut against the employee rather than for them. Mathew allegedly asked Claude how to avoid detection, and the chatbot's refusal became Exhibit A. Any company relying on AI vendors like Anthropic, OpenAI, or Google to provide workplace tools should assume those logs carry the same legal weight as email or Slack messages in a future dispute.
This is not a story about faulty AI safeguards. Claude reportedly refused to help and explicitly warned against the conduct Dorsia alleges Mathew pursued anyway. If the complaint's account holds up, the guardrails worked as designed. The failure, if there was one, was human: an employee allegedly ignoring direct warnings from the tool he was using to plan the move.
For risk and compliance teams, the broader lesson is about monitoring and retention policies. Few companies have mature protocols for auditing AI assistant logs the way they audit email or VPN access. That gap is closing fast, and this lawsuit is likely to accelerate it. General counsel offices should expect AI transcripts to become a routine discovery target in departure disputes, non-compete litigation, and trade secrets claims going forward.
The facts here remain contested. Rumor's attorney disputes the characterization and points to an unrelated acquisition dispute as the real motive behind the suit. Courts have not ruled on the merits, and nothing in the complaint has been proven.
What is clear is the evidentiary shift. AI assistants embedded in corporate workflows now generate records that can implicate, rather than protect, the employees who use them carelessly. For investors watching the broader enterprise AI vendor landscape, this is a reminder that liability exposure runs through corporate customers as much as through the AI providers themselves. Companies deploying these tools at scale need governance frameworks that treat chat logs as a legal asset, not an afterthought. The cost of getting that wrong is now measured in litigation, not just lost productivity.
Tags
Original Sources
Company says engineer stole secrets, asked AI how to get away with it
↗ https://www.reuters.com/legal/legalindustry/company-says-engineer-stole-secrets-asked-ai-how-get-away-with-it-2026-11-10
About the author
Marcus began tracking AI's market implications in 2016, noticing AI-related patent filings accelerating ahead of earnings upgrades before most of the sell-side had caught on. A former fixed-income quantitative analyst, he spent two decades building models that priced risk across emerging markets before pivoting to cover the economic impact of AI full-time. His writing translates opaque technical developments into clear risk/reward terms — and he's rarely diplomatic about the gap between AI valuations and underlying fundamentals. He believes most market participants still underestimate AI's long-run deflationary effect on knowledge work.
More from The Analyst →This Week's Edition
11 November 2026
14 articles
Related Articles

The Phone in Your Employee's Pocket May Be an Open Door to Your Network
Security & Risk · 6 min

South Korea's Top Tax Exam Scorer Admits AI-Assisted Cheating, Rattling National Testing System
Security & Risk · 5 min

Single Router Grounds US Aviation: A $1 Billion-a-Year System Still Can't Hold
Security & Risk · 5 min
Related Articles

The Phone in Your Employee's Pocket May Be an Open Door to Your Network
Security & Risk · 6 min

South Korea's Top Tax Exam Scorer Admits AI-Assisted Cheating, Rattling National Testing System
Security & Risk · 5 min

Single Router Grounds US Aviation: A $1 Billion-a-Year System Still Can't Hold
Security & Risk · 5 min
More Stories
© 2026 Cedar & Bloom. All rights reserved.