AI and Labor Relations: Why Your AI Rollout Is a Bargaining Event (Whether You Have a Union or Not)

Executive Summary

• AI deployment affecting wages, hours, or working conditions triggers a mandatory bargaining duty under the NLRA in unionized workplaces. Most mid-market GCs are treating AI rollouts as IT procurement decisions. The NLRB treats them as changes to terms and conditions of employment.
• The governing framework is NLRB General Counsel Memorandum 23-02 (Oct 31, 2022), which presumes a Section 7 violation when electronic surveillance and algorithmic-management practices, viewed as a whole, would tend to interfere with protected concerted activity. The memo remains the operative theory in pending cases.
• SAG-AFTRA’s May 2025 unfair labor practice charge against Llama Productions — alleging AI replacement without bargaining — is the first high-profile test of effects-bargaining doctrine applied to AI. Outcome shapes 2026–2027 enforcement posture.
• For non-union mid-market employers, AI rollout is itself an organizing trigger. Documented campaigns in 2025–2026 have cited algorithmic scheduling, surveillance, and AI discipline as the proximate grievance. The Labor Notes 2026 playbook names four specific strategies unions are using to organize around AI.
• The early-warning signs are observable before an organizing drive goes public. Collective framing of algorithmic complaints, group petitions for AI transparency, synchronized work slowdowns after rollout announcements, and external worker-center contact all predate card-signing.

Why This Is Different from a Normal Technology Rollout

An HRIS upgrade affects how HR works. An AI performance-management tool affects how every employee’s pay, promotion, and discipline is decided. The NLRA does not care that the vendor called it a “productivity assistant.” It cares about the effect on terms and conditions of employment.

Under current doctrine, an employer’s motivation for implementing AI is the decisive factor for determining bargaining duty. Labor-cost reduction triggers the duty. Entrepreneurial change of business direction may not, even when layoffs follow. The distinction is narrow and fact-specific, and the complaint-filing burden sits with the union — which means most mid-market GCs learn the rule after the charge lands.

Four AI deployments that have triggered or are likely to trigger a bargaining duty in a unionized workplace:

Source: NLRB GC Memo 23-02 (Oct 31, 2022); Proskauer Rose, “AI At Work: Safety And NLRA Best Practices For Employers” (2025).

The Non-Union Picture Is Not Safer

Mid-market CHROs tend to assume labor-relations risk applies only to unionized workforces. The NLRA Section 7 protects concerted activity by non-union employees as well. A group of employees who together protest an AI discipline tool are engaged in protected activity. Disciplining them for the protest is an unfair labor practice. Monitoring them to identify the organizers is potentially a Section 8(a)(1) violation under the GC 23-02 framework.

The Georgetown Center on Poverty and Inequality’s 2025 analysis documents employer use of AI to predict organizing behavior before it surfaces through conventional signals. The AFL-CIO’s April 2025 “Chatbots as the new frontline of union busting” report describes employers deploying automated anti-union messaging during campaigns. Both practices intersect with Section 7 and GC 23-02 in ways that a mid-market GC focused on IT security reviews would not flag.

Early-Warning Signs That AI Rollout Is Creating Organizing Conditions

These signals appeared in 2025–2026 organizing campaigns before any formal activity became visible. None is individually conclusive. Clustered, they are the pattern:

1. Employee complaints about algorithmic decisions framed collectively (“they’re doing this to us”) rather than individually.
2. Group petitions or open letters requesting AI transparency, data rights, or human review of algorithmic decisions.
3. Social media posts by employees naming specific AI tools used against them — particularly when posts reference coworker experiences.
4. Inquiries to HR about state privacy law rights (CCPA, Illinois BIPA, Colorado SB 205) following AI rollout. Often a precursor to NLRA-protected activity.
5. External contact with worker centers, union research staff, or employment lawyers. Visible in LinkedIn activity and sometimes in phone logs on employer devices.
6. Unexplained synchronized work slowdowns or absences in the week following an AI rollout announcement.

The State-Law Overlay Every Mid-Market GC Is Missing

A unionized employer in Colorado deploying AI for promotion decisions faces both NLRA effects-bargaining duty and Colorado SB 205’s high-risk AI notice and opt-out requirements (effective Feb 2026). A retailer in New York City faces NYC Local Law 144 bias-audit requirements on automated employment decision tools. These obligations run in parallel with the NLRA — not in place of it.

What the Unions Are Actually Doing in 2026

Labor Notes’ March 2026 playbook “Four Union Strategies to Fight on AI” names what is now on the bargaining table:

1. Advance-notice clauses (30–180 days) for any AI tool affecting bargaining-unit work.
2. Joint technology committees with pre-implementation review rights.
3. Worker data access rights — employees can see what data the AI uses about them.
4. Legislative push (California Labor Federation, AFL-CIO Workers First AI Summit) for predictive-AI restrictions in HR.

The CalMatters February 2026 reporting on union AI demands in Sacramento shows these are no longer theoretical — they are active 2026 contract-cycle demands at large employers. Mid-market employers with union relationships should expect the same demands in the next renewal.

Key Data Points

What This Means for Your Organization

If you have a union, your AI governance process should already route every AI deployment through a labor-relations review before procurement closes. If it doesn’t, you are budgeting the vendor cost but not the unfair-labor-practice risk, the delay from mid-deployment bargaining, or the injunctive relief risk if the deployment has already moved bargaining-unit work. Effects bargaining is the floor, not the ceiling — and “we didn’t know it was a mandatory subject” is not a defense.

If you don’t have a union, the calculation is different but the exposure is real. AI rollouts are the single fastest-growing category of Section 7 concerted-activity claims in 2025–2026. The mid-market companies that navigate this well share three practices: they disclose the AI tool and its data inputs before deployment, they preserve a human decision-maker for discipline and termination, and they train managers not to retaliate against the collective complaint patterns that GC 23-02 protects. The mid-market companies that get this wrong learn about the NLRA the same week they learn about the organizing campaign.

If the labor-relations dimension of your AI strategy hasn’t been stress-tested against the NLRA and the state-law overlay, that conversation is worth an hour. I’d welcome the discussion — brandon@brandonsneider.com.

Sources

• NLRB, “NLRB General Counsel Issues Memo on Unlawful Electronic Surveillance and Automated Management Practices,” Oct 31, 2022. https://www.nlrb.gov/news-outreach/news-story/nlrb-general-counsel-issues-memo-on-unlawful-electronic-surveillance-and — Credibility: HIGH (primary agency source).
• Proskauer Rose LLP, “AI At Work: Safety And NLRA Best Practices For Employers,” 2025. https://www.proskauer.com/pub/ai-at-work-safety-and-nlra-best-practices-for-employers — Credibility: HIGH (law firm synthesis of primary doctrine).
• Cooley LLP, “AI in the Workplace: US Legal Developments,” Sep 4, 2025. https://www.cooley.com/news/insight/2025/2025-09-04-ai-in-the-workplace-us-legal-developments — Credibility: HIGH.
• Baker McKenzie, “Navigating Labor’s Response to AI,” Jun 2025. https://www.bakermckenzie.com/en/insight/publications/2025/06/navigating-labors-response-to-ai — Credibility: HIGH.
• Labor Notes, “Four Union Strategies to Fight on A.I.,” Mar 2026. https://labornotes.org/2026/03/four-union-strategies-fight-ai — Credibility: MEDIUM (labor advocacy; useful for union bargaining posture).
• CalMatters, “Union leaders have a message for Newsom: Regulate AI if you want to be president,” Feb 2026. https://calmatters.org/economy/technology/2026/02/newsom-needs-more-ai-regulation-to-be-president-say-unions/ — Credibility: MEDIUM (news).
• Georgetown Journal on Poverty Law & Policy, “Labor Organizing and AI Surveillance in the Workplace,” 2025. https://www.law.georgetown.edu/poverty-journal/blog/labor-organizing-and-ai-surveillance-in-the-workplace/ — Credibility: MEDIUM (academic journal blog).
• AFL-CIO Union Label and Service Trades Department, “Chatbots: The New Frontline of Union Busting,” Apr 23, 2025. https://unionlabel.org/2025/04/23/chatbots-the-new-frontline-of-union-busting/ — Credibility: MEDIUM (labor advocacy source; useful for employer-tactic documentation).
• University of Chicago Law Review Online, “NLRA Protections for AI-Driven Layoffs?” 2025. https://lawreview.uchicago.edu/online-archive/nlra-protections-ai-driven-layoffs — Credibility: HIGH (academic).

Brandon Sneider | brandon@brandonsneider.com April 2026