Seth Kugler, a partner at Grellas Shah LLP, warns that AI agents create new trade secret risks, allowing employees to access or inadvertently expose proprietary information through personal devices. Companies should adopt strict AI policies, restrict data access on unmanaged devices, update confidentiality agreements, and act quickly when misuse is suspected.
The trade case I litigated in the Northern District of California shows a former senior marketing agency employee copying client lists and using stolen documents as training inputs to build an AI version of his former boss, seeking a shortcut to a competing business. The court found the defense’s claim that files were placed in a “quarantine folder” and no clients were poached insufficient, issuing a temporary restraining order and preliminary injunction, viewing the AI agent as an “existential threat.”
AI contamination cannot be undone; even if source files are scrubbed, the model retains the proprietary data, making it hard to prove the AI isn’t operating because of that information. Personal devices amplify risk: desktop AI agents can scan local hard drives, ingest folder paths, and train on employer data without special permission, violating confidentiality agreements. Thousands of employees may be breaching agreements by running consumer AI agents on personal laptops that store IP.
Source: Crunchbase News · Summarized by HeadlinesBriefing