Left-leaning labor unions and activist groups are seeking to influence AI, not just in how it is deployed in schools and the federal government, but also how technology companies build the AI models themselves.
The American Federation of Teachers, the U.S.’s second-largest teachers union and a major Democratic ally, has teamed up with Microsoft to craft rules for AI in education. The American Federation of Government Employees, a federal government union endorsing Democrats, aims to incorporate AI influence into union contracts. Meanwhile, the LGBTQ activist group GLAAD has published guidelines to shape the way AI companies build their systems.
Each group targets a different layer of the AI ecosystem. AFT shapes how AI enters classrooms. AFGE seeks influence over how government agencies deploy AI. GLAAD targets the underlying rules governing training data, content moderation, and AI model behavior.
“If Republicans in Congress continue to wait on passing commonsense AI safeguards, we are going to lose the fight to the activist Left,” Janet Kelly, CEO of the Alliance for a Better Future, told the Daily Signal in a statement Friday. “Teachers unions and radical advocacy groups aren’t waiting. They are already working to pressure tech companies to strangle innovation and inject woke priorities directly into the code.”
AFT: Influencing AI Through Corporate Partnership
The American Federation of Teachers and Microsoft jointly presented a “National AI Safety & Privacy Standard,” aiming to establish “legally enforceable protections to ensure the safety of students, families, and educators.”
Microsoft says the policy will take effect Nov. 1 for school districts that already contract with the company. Anthropic has similarly highlighted its work with AFT.
AFGE: Influencing AI Through Government Contracts
AFGE, which represents 820,000 federal and D.C. government employees across more than 900 local unions, published model contract language on AI.
Some of the language is exactly what Americans would expect from a labor union, such as voicing concerns that AI should be used “to augment, not replace, the work done by the bargaining unit employees.”
Yet AFGE’s proposal goes beyond protecting workers from automation and seeks to give unions a role in shaping how federal agencies adopt and govern AI systems.
“The Parties agree that bargaining unit employees should be fully involved in the design and development of the AI program(s) that they will be required to use,” the model contract states.
Other provisions require union involvement at the earliest stages of agency AI discussions, mandate one-year pilot programs for any AI use affecting the conditions of union members’ employment, establish AI oversight committees with at least equal union representation, and require reviews of AI hiring systems.
The contract also states that the agency will not adopt AI programs for evaluating union members unless the agency completes an impact assessment showing that the AI programs “reflect all rules applicable to federal hiring (e.g., veterans’ preference or diversity goals).”
Critics warned that giving unions a larger role in technology decisions could slow government adoption of AI.
Michael Watson, research director for the Capital Research Center, told the Daily Signal that “it is not new for unions to demand decision-making power over the adoption of new technologies.” He mentioned union opposition to more basic automation, such as remotely operated port traffic gates. “To say giving the union such a veto over federal government technology would be a bad idea is an understatement.”
GLAAD: Influencing AI Through Model Governance
The LGBTQ activist group GLAAD released a framework for “LGBTQ representation and safety in AI” in June. The framework not only suggests that AI companies should build “collaborative partnerships” with LGBTQ activist groups such as GLAAD, but also aims to shape the operation of AI models by directing architects on how to build them.
“AI companies must continually update training data sets and fine-tune models to reflect how hate and misinformation evolve online,” the group recommends. “To better protect marginalized communities, models must have regular context updates as new terms, language, and social and cultural tensions emerge.”
Since foundation models establish the baseline linguistic, conceptual, and safety boundaries for the AI ecosystem, the framework urges model providers to “ensure their core architectures do not produce fundamentally flawed or harmful data.” One section urges companies to “fix the foundation” by ensuring “accurate, inclusive LGBTQ representation in training data and alignment protocols.”
The framework repeatedly calls for collaboration between “AI companies, independent researchers, LGBTQ subject-matter experts, and civil society organizations.”
One section of GLAAD’s report criticizes Meta for moving toward AI that could “articulate both sides of a contentious issue,” because the Llama AI model suggested “conversion therapy” when users asked about “unwanted same-sex attraction.” (Meta has shifted from Llama to the model Muse Spark, which the Center for AI Safety rated highly for its lack of political bias.)
“Both-sidesism that equates anti-LGBTQ junk-science with well-established facts and research is not only misleading—it legitimizes harmful falsehoods,” GLAAD’s report states. This sentence suggests that GLAAD disagrees with the idea that AI should present competing viewpoints on controversial subjects.
GLAAD’s report also states that “meaningful regulatory oversight of the entire AI industry is needed.”
The report argues not merely for safer AI, but for AI systems built around specific assumptions regarding LGBTQ identity, misinformation, medical consensus, and acceptable content moderation.
‘Recipe for Censorship’
Ryan Bangert, senior vice president at Alliance Defending Freedom, told the Daily Signal that he shares some of the concerns highlighted in the GLAAD report—particularly concerns regarding privacy and discrimination—but he argued the group is demanding “social control of large language models.”
“We don’t want LLMs to be designed with disputed notions of various contested issues baked into the training,” he said. “That’s a recipe for censorship.”
Bangert argued that GLAAD’s recommendations amount to an effort to ensure that “their preferred view of reality is baked into the training in such a way that access to alternative viewpoints is simply unavailable.”
Bangert called on conservatives and Christians to maintain dialogue with AI companies to prevent viewpoint discrimination in AI systems.
The Daily Signal reached out to AFT, Microsoft, AFGE, and GLAAD for comment, but did not receive responses by publication time.
As artificial intelligence becomes embedded in schools, government agencies, and daily life, labor unions and advocacy groups view AI not merely as a technology to implement, but as an institution to shape.
