Ex-employee “turned into an AI digital human to keep working” sparks controversy; company says it obtained the person’s consent for training

Allegations and company response

A widely circulated report has sparked debate in China after it was claimed that a technology company used an ex-employee’s voice and likeness to create an AI “digital human” that continued performing parts of that person’s job. It has been reported that the company told regulators and the public it obtained the individual’s consent to use their data for training the model. Reportedly, the digital twin has been deployed in customer-service and content-generation roles, raising questions about consent, labour rights and the limits of synthetic identity.

Why the story matters

Why does a single case matter to readers outside China? Because this incident touches on universal tensions: automation versus employment, the commercial reuse of personal data, and the legal status of digital replicas. In China, the Personal Information Protection Law (PIPL) and recent draft rules on generative AI set a backdrop of tightening oversight, but enforcement and gray areas remain. Companies argue commercial flexibility; critics point to power imbalances—can consent given under employment pressure ever be truly voluntary?

Broader tech and geopolitical context

The controversy also arrives amid global scrutiny of AI training practices and growing cross-border friction over data flows and export controls. For Chinese tech firms operating internationally, reputational damage can compound regulatory risk, especially as Western policymakers push for stricter rules on synthetic media and data provenance. It has been reported that industry groups in China are watching the case closely, fearing it could prompt faster rule-making or consumer backlash.

What’s next

Regulators and industry bodies now face pressure to clarify consent standards, record-keeping and the rights of individuals whose images and voices are used to train models. Will new guidance require explicit, revocable consent with documented terms? Or will companies continue to test the boundaries as business models evolve? Either way, the episode highlights a fast-moving ethical frontier at the intersection of labour, privacy and generative AI.