Scale AI sues former employee Eugene Ling and his current employer, Mercor, one of Scale's key competitors, for allegedly stealing 100+ confidential documents
Context & Ripple Effects
The suit follows a period in which Scale was defending the handling and separation of customer information, including its statement that Meta would not receive access to customers’ confidential data. A subsequent report that training documents for major customers were accessible through shared links put information controls at the center of Scale’s competitive posture.
The allegation also arrives amid a broader run of employer-versus-departing-employee disputes in AI: Meta’s suit over documents allegedly taken to an AI startup and xAI’s trade-secret case involving a move to OpenAI show companies using litigation to contest how talent and proprietary materials move between rivals.
First-order effects
- Scale is seeking to stop or remedy Mercor’s alleged use of more than 100 confidential documents, putting the former employee and Mercor under immediate legal and reputational pressure.
- The case makes Scale’s document controls and treatment of customer- and training-related information more salient, particularly after reporting on confidential training materials accessible through shared links.
Second-order effects
- Mercor may need to demonstrate separation from the disputed materials and strengthen onboarding, access, and provenance controls; similar rivals have reason to review how they handle hires from direct competitors.
- Customers whose work depends on protected data may apply more scrutiny to vendors’ access governance and contractual safeguards, rather than treating data-labeling and AI-training operations as interchangeable services.
Third-order effects
- If these disputes continue, AI services competition could shift toward formalized trade-secret boundaries around workflows, datasets, and customer context, making employee mobility more legally costly without determining the merits of any individual claim.
- The pattern favors vendors able to evidence data provenance and compartmentalized access; it may also make governance a differentiator alongside model-training capacity and labor supply.
The trend: This is one instance of the talent-to-trade-secret transition, in which AI competitors increasingly treat employee departures as potential transfers of operational IP.