AI Digital Twins of Laid-Off Employees: A New Frontier in Workplace Ethics and Labor Law

A heated discussion on Chinese Q&A platform Zhihu has sparked a debate about companies using AI "digital twins" of laid-off employees — AI systems trained on former workers' communication styles, code patterns, and work habits that continue to function in their place after they've been terminated.

The Practice

According to the discussion, some companies are:

• Training AI models on departing employees' data — Communication patterns, coding styles, work preferences
• Deploying these "digital twins" post-termination — The AI continues responding to messages, writing code, and handling tasks
• Reducing headcount while maintaining output — One employee's departure doesn't mean one employee's worth of work stops

Key Questions

Legal Implications

Several legal issues arise:

• Data ownership — Who owns the patterns and knowledge encoded in the AI?
• Privacy rights — Was the employee's consent obtained for their data to be used this way?
• Intellectual property — If the AI writes code based on a specific programmer's style, who owns the output?
• Contractual obligations — Employment contracts rarely address post-termination AI replication

Ethical Considerations

• Deception — Is it ethical to have an AI impersonate a human in workplace communications?
• Dignity — Does this practice respect the departing employee's professional identity?
• Exploitation — The employee trained the AI during employment but receives no benefit from its continued use
• Transparency — Should colleagues be informed they're interacting with an AI?

Labor Market Impact

If widely adopted, digital twin practices could:

• Reduce bargaining power — Employees become more easily replaceable
• Accelerate automation — Companies invest more in capturing and replicating human knowledge
• Create new job categories — "AI human mimicry engineers" or "digital identity consultants"
• Widen inequality — Knowledge workers most at risk; manual laborers less so

Regulatory Response

This practice falls into a regulatory gray area:

• China's Personal Information Protection Law (PIPL) may apply
• EU AI Act could regulate similar practices in Europe
• US labor law has not yet addressed AI replacement of specific individuals
• China's Cybersecurity Law has provisions about data processing and consent

The Bigger Picture

This is part of a broader trend of AI replacing not just tasks, but specific people:

• Companies like Google and Amazon have already replaced human workers with AI in customer service
• The "digital twin" concept takes this further by creating AI replicas of specific individuals
• As AI capabilities improve, the line between human and AI-generated work will increasingly blur

What Should Be Done?

Possible regulatory frameworks could require:

1. Explicit consent before training AI on employee data
2. Transparency obligations — Informing stakeholders when AI is impersonating humans
3. Compensation frameworks — Employees compensated for their data's post-employment value
4. Usage limitations — Restrictions on how long and for what purposes digital twins can be used

Source: Zhihu discussion, labor law analysis