Education International

The Dialogue would be a success if it delivered a principled, action-oriented and enforceable framework that anchors AI governance in human rights, democracy, climate justice and the public interest. It must account for national governance contexts while setting up a global framework for AI governance, given the cross‑border operation of AI companies and products. From an education and teacher perspective, success includes explicit recognition that AI is not neutral, but a socio-technical system reshaping power, work, knowledge production, and the public good. The impact of AI on student and teacher learning, including cognitive and social-emotional development and well-being remains insufficiently understood and carries significant risks. The late realization of the negative impact of social media and increasing calls for bans underscores the need for a more cautious, preventative approach to the use of AI in education institutions. The Dialogue should establish a clear normative consensus that AI governance must enhance human capabilities, following human-centered and humans-in-the-lead principles. AI in education and research must strengthen and not replace or deprofessionalise educators, researchers, and education support personnel recognising education as a fundamentally relational endeavour with the student-teacher relationship at its core. The Dialogue will succeed only if it meaningfully consults workers and key stakeholders, including education unions and makes their involvement key and center in any recommendations on developing regulatory frameworks through structured policy and social dialogue, in line with the Santiago Consensus adopted at the UNESCO World Summit on Teachers. This can ensure AI tools are safe, pedagogically sound, support quality and equitable education and respect working conditions and professional autonomy. The development of regulatory frameworks needs to be resourced to establish expert panels to assess AI tools against agreed quality standards. Reference documents: ETUCE policy paper on AI: https://ziply.pk/35SJos EI global resolution: https://ziply.pk/7miAZg AI conference conclusions: https://ziply.pk/H2fnaB

• Safe, secure and trustworthy AI
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Protection and promotion of human rights
• Transparency, accountability, and human oversight

The rationale is provided through other responses. In particular, in the first and third question.

A critical cross-cutting issue insufficiently addressed is the impact of AI on professional autonomy, professional rights and working conditions. In education systems worldwide, AI is increasingly deployed to standardise pedagogical practices, automate assessment, and intensify monitoring and evaluation of educators. This raises fundamental governance questions about professional autonomy and academic freedom, accountability, and the protection of teachers' rights. In addition, the working conditions of data workers in the production of AI systems is a key concern across sectors. Another missing theme is dialogue and meaningful consultation. Any framework on AI Governance must refer to the importance of institutionalized policy and social dialogue. In the education sector, governments have to ensure that the teaching profession is represented in all decision-making processes regarding the use of AI in the education and research sector. The Santiago Consensus mirrors this central concern of ensuring teachers are in the lead in relation to new technologies. It recognises that education technologies should be "co-created and utilized by teachers to amplify the relational and affective dimension of learning without replacing the creativity, core competencies, and pedagogical knowledge of teachers." (UNESCO and Government of Chile, 2025, article e). Only through policy and social dialogue and collective bargaining can we ensure that the cognitive and social-emotional needs of students are safeguarded and the professional rights of teachers are protected. often without meaningful social dialogue, negotiation, or consent. The increasing evidence on the cognitive decline and atrophy as well as emotional (e.g. anthropomorphic AI) and social development impacts of AI on teachers and students across all education levels is a key concern also beyond education that should be considered in any dialogue an AI governance. There is also an urgent need to address the environmental footprint of AI, including energy consumption and resource extraction across the technology supply chain. These impacts disproportionately affect communities already facing climate, social, and economic vulnerabilities. AI governance must therefore be aligned with global commitments to climate justice, sustainability, and just transition. Ultimately, the responsible use of AI technologies in education and beyond should reflect a holistic commitment to sustainability in every aspect of society.

The education sector is a major and fast-growing target for the AI market often treated as a space for innovation, experimentation rather than a public good and human right. Governance gaps mean that AI is advancing faster than democratic control, social dialogue with education unions, and evidence-based policy in education. AI systems are being introduced into classrooms, assessment, student monitoring, and administrative decision‑making without clear, enforceable safeguards. Existing legislation (e.g. EU AI Act, data protection rules) is fragmented or insufficiently adapted to the realities of teaching and learning which creates legal uncertainties. In the absence of strong regulation, education risks becoming an experimental testing ground for edtech products, with limited evaluation of their pedagogical value, accuracy, bias, or long‑term societal impact. At same time without regulation there is no accountability system leaving the responsibility with teachers that do not have designed the AI systems they are asked to use. Students are a uniquely vulnerable group. Governance gaps around safety, transparency and accountability expose them to the risks of commercial exploitation, automated decision-making that may shape educational pathways and life chances for years to come, they may impact their cognitive as well as social and emotional development. Further, clear regulations and guardrails for the use of AI in classrooms and schools is absent. There are increasing findings on the cognitive atrophy and cognitive decline as well as emotional and social development issues related to the use of AI by young students but also what this does with teachers in terms of their professional learning, motivation and development. It is positive to see that regulation is being discussed in the EU as well as in the African union and also that education is identified as a high-risk sector.

The AI Dialogue can play a vital role by providing a structured, inclusive space for dialogue across regions, bringing together governments, educators' unions, civil society, and international organisations. It can help surface real‑world impacts of AI in education, particularly from the perspective of educators and learners, whose voices are often omitted from technical or market‑driven discussions and help to make best use of AI in the interest of humanity. AI dialogue cannot be a substitute for binding regulation, but it can be a critical mechanism and platform to coordinate across borders to help bridge policy gaps, elevate education-specific risks and build coherent international governance rooted in human rights, democracy and the public interest.

Education International as the Global Union Federation bringing together organisations of teachers and other education employees worldwide and representing more than 33 million teachers and education support personnel in 180 countries and territories has been a key partner in contributing to as well as developing and implementing policies related to education and research – either through dialogue with our global leadership through the work of member organisations at the national and regional level. In addition, to our global secretariate that coordinates dialogue with international organisations, we have regional offices in all regions of the world engaging with key partners and initiatives as the ones listed below: UNESCO OECD (Opportunities, guidelines and guardrails for effective and equitable use of AI in education) African Union International Summit of the Teaching Profession International Labour Organisation UNICEF GPE From the perspective of the education sector, collective bargaining, social and policy dialogue that often takes place at the national level as well as in cross-regional contexts like the EU, has been proven to be very effective in ensuring quality development, implementation and evaluation of education policies including in relation to learning and students as well as working conditions of teachers. Consultation with key stakeholders representing workers in various sectors including Education International.

Governments and public authorities: regulatory expertise and translating dialogue outcomes into binding policies, public procurement rules and enforcement Education unions (and other sectoral unions): essential contributors bringing frontline experiences on education from a holistic and professional perspective including its impact on pedagogy, student well-being, equity, working conditions, professional autonomy and academic freedom. Their involvement ensures that AI governance reflects real-world educational practices and needs rather than abstract technical and and profit-driven considerations. Students and civil society actors: critical perspectives on equity, inclusion, data protection, and long‑term societal consequences, particularly for children and young people International organisations and experts: can support coherence across regions, share comparative evidence, and connect the Dialogue to existing global frameworks. Technology providers : may offer technical insights, but their participation must be transparent and strictly balanced to prevent regulatory capture and ensure AI governenace serves the public interest and not commercial priorities. Minority and marginalized communities: lived experiences and identify real life risks of bias, exclusion and discrimination in AI systems. Needed for inclusive design and accessibility. Format and structure: Plenary sessions for cross-sectoral discussions Sector-specific working groups with education being recognised as a high-risk sector Clear rules on participation, transparency, representation Outcome-focused process, resulting in concrete recommendations Continuous monitoring and follow-up

Broadly speaking, the education sector has often been left out of discussion on AI use for teaching and learning or national unions have had to lobby significantly to get a voice at key consultations. That needs to change.

Sector‑specific conversations are essential to ensure the Dialogue moves beyond vague principles and provides concrete support for complex realities. Effective regulation and governance can only be achieved this way.

Effective AI governance requires clear regulatory frameworks that protect fundamental rights, ensure transparency, accountability and human oversight, and prevent harm to learners and education professionals. AI-enabled technologies used in education must be approached with caution, as their deployment can have far-reaching consequences for pedagogical practices, professional autonomy, working conditions, learning outcomes as well as social- emotional well-being. Education must be recognised as a high-risk sector, and AI systems used for teaching, learning, assessment, monitoring, profiling, or management should be subject to strict safeguards and prior risk assessments. Decisions on the introduction and use of AI-enabled technologies in education must be evidence-based, pedagogically justified, and aligned with public education values. Social dialogue and collective bargaining are essential governance tools for addressing the impact of AI in education. Teachers, researchers, and education support personnel, through their trade unions, must be involved at all stages of the design, procurement, implementation, monitoring, and evaluation of AI systems. This participation is necessary to prevent deprofessionalisation, work intensification, algorithmic management, and surveillance. AI governance must also ensure full compliance with data-protection and privacy legislation, including limitations on data collection and processing, clear purpose specification, and safeguards against the commercial exploitation of education data. Public authorities should favour publicly governed, transparent and accountable digital infrastructures over opaque, profit-driven platforms. Finally, comprehensive training and professional development are required to support educators in understanding the ethical, legal, social, and pedagogical implications of AI-enabled technologies. AI must be used to support and not to replace the relational, human-centred nature of education, respecting academic freedom, inclusion, accessibility, and equity.