Cyber Worker! Former employees are turned into AI clones to continue their roles

【This article was published on 04/07 by @Xiaonaixinren, the author of Xiaoheihe; reprints must indicate the source!】

People have left, but the class is still in session; they resign, but the work is still getting done.

Recently, a bold move by a game media company in Shandong has drawn attention. A former HR specialist—using data from their work documents, chat logs, communication tone, and other information during their employment—was trained into a dedicated AI digital double. Even though the person had already completed their resignation paperwork, this “cyber coworker” still remains on the job, handling basic tasks such as consultation replies, PPT creation, and spreadsheet organization.

According to a disclosure by a currently employed staff member at the company involved, this AI double experiment was not a forced measure. The departing employee was informed and agreed to it in advance, and even proactively provided an avatar and work materials, joking that it was “pretty fun.” At present, the AI double is only in an internal testing stage and has not yet been deployed for external use. Its capabilities are also not considered mature; coworkers have mocked it as “Doubao for working people—somewhat slow, it can only handle simple instructions.”

The company states that this effort is not intended to replace in-service employees, but rather is a frontier attempt. The goal is to hand repetitive, basic tasks to AI, freeing up human resources to do more creative work. In the future, the company also plans to develop humanoid robotic staff to handle office reception, appointment guidance, and similar matters.

In response to this case, the legal community has clearly issued risk warnings. Under the Personal Information Protection Law, employees’ chat records, work habits, and the like are personal information, and private content is even sensitive personal information. The Interim Measures for the Administration of Generative AI Services also stipulate that training AI using personal information requires obtaining the person’s consent.

Even if the employee volunteers, the company may not use the data beyond the permitted scope. If personal information is used without permission or information is abused, it will infringe on individuals’ rights and interests in personal information. In serious circumstances, it may also constitute the crime of infringing on citizens’ personal information, with a maximum penalty of up to seven years of fixed-term imprisonment.