Scale AI Sues Former Employee and Mercor for Alleged Customer Theft

Scale AI Files Lawsuit Against Former Employee and Rival Mercor

Scale AI, a company specializing in data preparation for AI model training, initiated legal proceedings on Wednesday against a former sales employee and its competitor, Mercor. The lawsuit alleges the employee, subsequently employed by Mercor, illicitly obtained “over 100 confidential documents” detailing Scale’s customer strategies and other proprietary information, as reviewed by TechCrunch.

Details of the Legal Action

The suit targets Mercor for misappropriation of trade secrets and Eugene Ling, the former employee, for breach of contract. Furthermore, the claim asserts that Ling attempted to solicit Mercor’s services with one of Scale’s key clients prior to his official departure. This client is referred to as “Customer A” within the legal documents.

Mercor's Response

Surya Midha, a co-founder of Mercor, firmly denies that his company utilized any data originating from Scale. He acknowledges, however, that Ling may have possessed such data.

“Mercor has welcomed numerous individuals previously employed by Scale, but we have no desire to leverage Scale’s trade secrets and are deliberately operating our business differently,” Midha stated in an emailed communication to TechCrunch. “Ling informed us of the presence of older documents within his personal Google Drive, which we have never accessed and are currently investigating.”

Midha further indicated, “We contacted Scale six days ago, offering to facilitate the destruction of these files by Ling or to reach an alternative resolution, and we are presently awaiting their response.”

Allegations Regarding Customer A

Scale contends that the documents in question contained specific data that would enable Mercor to effectively serve Customer A, alongside several other of Scale’s most significant clients.

Demands and Ling's Statement

Scale requested that Mercor provide a comprehensive list of files contained within the Google Drive and prevent Ling from engaging with Customer A. The lawsuit claims Mercor declined this request. Ling initially did not respond to inquiries from TechCrunch, but later posted on X (formerly Twitter): “Just learned I’m being sued by Scale. Last month, I transitioned from Scale to Mercor. I understand this caused frustration within my previous team, and I regret that.”

Ling continued, “Upon Scale’s inquiry regarding files stored in my personal drive, I offered to delete them. However, Scale requested I refrain from any action, and I am currently awaiting guidance on resolution. I have not utilized these files in my current role. It appears Scale intends to pursue legal action, which is their prerogative. I want to emphasize that there was no malicious intent involved. I apologize to my new team at Mercor for the complications this has caused.”

The Significance of Customer A

The lawsuit provides limited details regarding the identity of Customer A. It does reveal that securing this client would represent a contract “valued at millions of dollars” for Mercor.

Implications and Competitive Landscape

Regardless of the specifics of this legal dispute, it underscores Scale’s concern regarding the competitive threat posed by Mercor, prompting the pursuit of legal remedies. As previously reported by TechCrunch, despite Meta’s substantial investment in Scale, TBD Labs – Meta’s division focused on AI superintelligence – continues to utilize Mercor and other LLM data training service providers.

Mercor's Growing Presence

Mercor is gaining prominence in the LLM training sector due to its practice of employing content specialists, frequently holding PhDs, to refine LLM data within their respective fields of expertise.

Recent Developments

In June, Scale announced a $14.3 billion investment from Meta for a 49% stake in the company, accompanied by the recruitment of its founder. Following this announcement, several of Scale AI’s major data customers, who are competitors to Meta’s initiatives, reportedly severed their ties with the company.

Updated with comments on social media from Eugene Ling.