American Federation of Teachers

The American Federation of Teachers (AFT) is the United States' second-largest teachers' union and one of the country's largest labour organisations overall, representing 1.8 million members — K–12 educators, higher-education faculty, school-support staff, government employees, and healthcare workers — across more than 3,000 local affiliates nationwide. An AFL–CIO affiliate since its founding, the AFT has positioned itself as the most assertive institutional voice demanding that artificial intelligence be treated as a mandatory subject of collective bargaining in American schools — while simultaneously operating the largest union-run AI-training initiative in US education history. This dual posture — demanding worker-driven guardrails in contract negotiations while channelling $23 million in Big Tech investment into educator AI-literacy infrastructure — makes the AFT the most complex institutional actor on AI in the US education sector.

Founding and structure

The AFT was founded on 15 April 1916 in Chicago, when eight local teaching unions convened and petitioned the American Federation of Labor for a charter. AFL President Samuel Gompers granted that charter on 9 May 1916, creating the first explicitly union-model national organisation for teachers — organised labour rather than the professional-association model that the older National Education Association represented. The labour model's defining early feature was the exclusion of school administrators from membership: the AFT was for teachers, not for the administrators who managed them. The AFL affiliate structure tied the AFT to the house of labour and to its traditions of collective bargaining and strike action, distinguishing it from the NEA's preference for professional-advocacy mechanisms.

Today the AFT's membership spans five divisions: pre-K through twelfth-grade teachers; paraprofessionals and other school-related personnel; higher-education faculty and professional staff; federal, state, and local government employees; and nurses and other healthcare professionals. The approximately 80,000 early-childhood educators and nearly 250,000 retirees affiliated with the AFT are in addition to the core 1.8 million. The union elects its national officers at a biennial convention at which delegates also set union-wide policy. Randi Weingarten has served as AFT president since her election on 14 July 2008, succeeding Edward McElroy; Weingarten had previously led the United Federation of Teachers — the AFT's New York City local and its largest single affiliate — and is the most recognisable individual voice in the US teachers'-union movement.

AI as mandatory subject of bargaining — the 2024 resolution

At the AFT's 2024 biennial convention in Houston, delegates passed a resolution on artificial intelligence that established the union's core labour-law stance: the impact of AI in the workplace is a mandatory subject of bargaining, and AFT will develop contract language, policies, and practices to support affiliates at the bargaining table. The resolution simultaneously set the union's substantive AI positions: the AFT endorses responsible and principled AI development and deployment; opposes any AI application that leads to worker displacement, infringes collective-bargaining rights, perpetuates systemic biases, or widens the digital divide; and demands that transparency, accountability, corporate responsibility, respect for intellectual property, and worker privacy be treated as non-negotiable baseline conditions of AI deployment in AFT-represented workplaces. The resolution's framing of AI as a mandatory bargaining subject — rather than a management prerogative — was the most consequential labour-law position a major US union had taken on AI at that point in the US education sector: it converted the "we have concerns" posture of an advocacy statement into a "this is negotiable, and we will negotiate it" posture of a union preparing for the table.

"Commonsense Guardrails" and early organising (2024)

In June 2024, the AFT announced "Commonsense Guardrails for Using Advanced Technology in Schools" — a framework report compiled by classroom practitioners addressing six dimensions of AI risk in schools: accuracy and reliability, data privacy, equity and bias, transparency, intellectual property, and professional autonomy. The report insisted that traditional educational goals — critical thinking, literacy, civic engagement — must remain primary and that technology cannot be allowed to replace direct in-person educator–student relationships. The AFT simultaneously committed $200,000 to fund AI governance and integration projects in eleven school districts across the country, with the explicit goal of generating replicable models for incorporating and regulating AI that bargaining units in other districts could adopt. The union also launched the AI Educator Brain, a free resource library hosted on the Share My Lesson platform, as an initial professional-development channel.

National Academy for AI Instruction (2025)

In July 2025, the AFT announced the National Academy for AI Instruction — a $23 million initiative to provide free AI training and curriculum to all 1.8 million AFT members, beginning with K–12 educators. The launch partners were the United Federation of Teachers as co-convener; Microsoft as lead partner ($12.5 million over five years); OpenAI as founding partner ($8 million in direct funding plus $2 million in technical resources); and Anthropic ($500,000 in first-year support). The Academy is anchored at a purpose-built Manhattan facility and will operate through a combination of workshops, online courses, and hands-on training sessions designed to train 400,000 educators in the first phase. The AFT framed the initiative explicitly as putting educators "in the driver's seat" on technology in classrooms — the union as the governing actor, with tech-company resources funding its educator-led programme rather than the tech companies setting the curriculum.

The partnerships drew criticism. Critics noted that Weingarten was accepting large sums from OpenAI and Microsoft — companies that had been major targets of the AI guardrails the AFT had called for at its 2024 convention — and that the arrangement could compromise the AFT's ability to bargain adversarially against AI vendors in education. The AFT's position in response was that the Academy's curriculum was union-designed and educator-controlled, and that member-driven AI literacy was itself a precondition for effective bargaining: workers who understand the technology can negotiate smarter contract provisions than workers who do not.

Contract fights and strike action

The AFT's mandatory-bargaining stance has translated into concrete contract disputes. At the University of Illinois Springfield, faculty represented by the AFT struck in part over management's refusal to "establish meaningful protections around the use of artificial intelligence in the classroom" — the first reported US faculty strike in which AI governance was named as a central issue. The University Council-AFT, representing non-tenure-track faculty at the University of California system, proposed a new collective-bargaining article on Technology in 2025 negotiations, with members from multiple campuses documenting their concerns about AI tools in instructional settings. Weingarten has publicly framed future teacher strikes as likely to centre on intellectual property, employee data privacy, and algorithmic bias — the set of AI-related issues the 2024 resolution identified as mandatory subjects — rather than the traditional compensation and class-size issues that anchored prior rounds of teacher bargaining.

"Devices Down, Eyes Up, Hands-On" — the 2026 pivot (May 2026)

In May 2026, Weingarten delivered a major education address at the National Press Club in Washington, D.C., unveiling a 10-point plan titled "Devices Down, Eyes Up, Hands-On: 10 Points to Boost Teaching and Learning in the AI Era." The plan called for: a screen ban — including online assessments — for students in pre-K through second grade; no student-facing AI in elementary schools, on the grounds that building foundational skills, relationship capacity, and persistence requires human-centred development; a total ban on social-companion chatbots for students under sixteen; an enforceable privacy standard for AI use in schools; a technology tax on AI companies to compensate the public for AI's social costs; and deep classroom-relevant research before any new AI tool reaches students. Microsoft, OpenAI, and Anthropic each endorsed the overarching tenets of the proposed standard — the same companies whose money finances the National Academy for AI Instruction. The simultaneous positioning of the AFT as a Big Tech funding partner and a Big Tech regulator marks the central tension in the union's AI posture as of mid-2026.