Can Good-Faith Bargaining Lead to Better AI Implementation - The National Law Review

TEXT START: Like any form of workplace automation, AI is accelerating employee concerns over displacement, surveillance, bias, productivity measuring and more.

The Dissection

This is an employer-side transition-management memo disguised as a discussion of worker protection. It converts displacement, surveillance, bias, discipline, staffing, and patient safety into negotiable implementation variables: notice, timing, safeguards, use cases, and employee buy-in. Its objective is smoother AI deployment with less legal friction and resistance.

The Core Fallacy

It confuses participation in implementation with control over the economic outcome. Under the Discontinuity Thesis, bargaining may slow or shape substitution, but it cannot defeat P1, P2, or P3. If AI delivers superior cost and performance, competitive pressure forces adoption. Transparency makes the machinery visible; it does not make workers owners of it. Buy-in turns opposition into implementation capacity. Even “consistent discipline,” presented as a benefit, is an automated management function that can make human supervision and labor more disposable.

Hidden Assumptions

• AI will mainly enhance existing jobs rather than eliminate economically necessary labor.
• Good-faith bargaining can preserve staffing, security, and human relevance instead of merely negotiating the terms of reduction.
• Initial agreements can govern rapidly expanding use cases.
• Employers will sacrifice enough efficiency to preserve jobs when competitors automate.
• Training, human review, and procedural safeguards will remain economically scalable.
• Bias, transparency, and safety are the central problems, while ownership of the productivity gains remains unexamined.
• Unionized workers’ legal leverage represents a general solution, although the proposed protections apply only in the unionized context.

Social Function

Classification: transition management, ideological anesthetic, and partial truth.

The partial truth is real: bargaining can improve immediate safeguards, healthcare deployment, data use, discipline rules, and legal compliance. But the text recasts a distributional conflict—who owns the gains and who loses income—as a communication and governance problem. “Involve workers early” is useful operational advice and effective resistance management. It is a lubricant for automation, not a brake on its structural endpoint.

The Verdict

Good-faith bargaining can produce safer, slower, more legitimate AI implementation. It cannot preserve mass productive participation once AI becomes cheaper and better across cognitive work. Unions may extract disclosure, human review, staffing floors, severance, or transition payments—a toll on the machinery. They cannot make the machinery need the workforce. This is competent hospice protocol for the wage-consumption circuit, not a cure.