Top 20 ETIH stories of the year: Counting down 2025's biggest moments in EdTech - part one
From AI education pledges and workforce investment to mounting questions around assessment, safety, and trust, these were the stories that defined EdTech’s most revealing year yet.
2025 was the year EdTech stopped dealing in hypotheticals. As artificial intelligence moved deeper into classrooms, institutions, and policy, the sector was forced to respond to real pressure points, from skills shortages and academic integrity to data security and student safety. This two-part series looks back at the stories that captured those shifts.
In part one, we revisit the stories ranked 20 to 11, the moments that set the direction of travel and shaped the year before the biggest headlines took hold. The developments set the tone for the year and revealed how governments, universities, schools, and technology providers began grappling with what AI-driven education actually looks like in practice. Part two, covering the top ten stories, will follow next week.
20. White House gathers Big Tech to back AI education with funding, training, and tools
The White House convening on AI education marked one of the earliest signals of how seriously governments planned to lean on industry in 2025. With Amazon, Google, Anthropic, Brainly, and others signing the Pledge to America’s Youth, the focus shifted from abstract commitments to concrete promises around funding, training, and classroom tools.
What stood out was not just the scale of the pledges but their coordination. Amazon committed to training millions of learners and supporting educators with AI curricula, while Google confirmed nationwide access to Gemini for Education and outlined significant investment in AI literacy and digital wellbeing. Anthropic and Brainly focused on curriculum access and classroom integration, with an emphasis on avoiding lock-in and expanding reach to underserved communities.
The moment reflected a broader pattern that would repeat throughout the year. Governments increasingly positioned AI readiness as a national priority, while technology companies stepped into roles traditionally held by public institutions, shaping how students and educators would encounter AI in practice. The challenge, made clear even at this early stage, was whether coordination and accountability could keep pace with ambition as these initiatives moved from pledge to implementation.
19. Google DeepMind launches AI for Math Initiative with five research institutions
Google DeepMind’s AI for Math Initiative signaled how far artificial intelligence had moved beyond classroom tools and into the foundations of academic research. Partnering with institutions including Imperial College London, the Institute for Advanced Study, IHES, UC Berkeley’s Simons Institute, and the Tata Institute of Fundamental Research, the initiative focused on using AI to tackle mathematical problems that have traditionally resisted human approaches.
Rather than positioning AI as a replacement for expertise, the project framed it as a collaborator. DeepMind’s work with systems such as AlphaGeometry, AlphaProof, and Gemini Deep Think demonstrated how machine reasoning could surface new patterns, test conjectures at scale, and open routes into problems previously considered intractable. The emphasis was on feedback loops between researchers and models, with AI supporting exploration rather than dictating outcomes.
The initiative stood apart because it placed AI at the edge of human reasoning rather than at the point of delivery. By focusing on unsolved mathematical problems and abstract discovery, DeepMind positioned AI not as an efficiency tool but as a research partner operating in unfamiliar territory.
18. Curtin University drops Turnitin’s AI detector from 2026 as reliability debate grows
Curtin University’s decision to disable Turnitin’s AI writing detection feature from January 1, 2026 landed as a clear signal about where institutional trust is starting to fray. Originality checking will remain in place, but AI detection will not. Curtin framed the change as a move toward greater clarity and fairness in assessment, rather than a retreat from academic integrity.
Students will continue submitting work through Turnitin, with assignments still reviewed for consistency and integrity. What changes is the evidence base. AI detection flags will no longer be used, and the university has pointed students toward alternative tools such as Curtin Grammarly and the Authorship function, alongside ongoing review of assessment design to better reflect real-world use of AI.
The timing pushed this story into sharper focus. Curtin’s announcement came amid growing academic scrutiny of AI detection tools, including calls for independent testing, transparency around false positives, and clearer accountability when tools influence high-stakes decisions. While Turnitin emphasized that human judgment remains central, Curtin’s move reflected a broader recalibration underway across higher education.
17. Intelsat and MaxIQ Space expand Africa Space STEM Program across four countries
The expansion of the Africa Space STEM Program stood out in 2025 for what it did quietly rather than loudly. Intelsat and MaxIQ Space confirmed the program would reach more schools across Kenya, Nigeria, South Africa, and Senegal, extending a long-running initiative that prioritizes sustained access to STEM education rather than short-term pilots.
Now in its fifth year, the program moved beyond virtual engagement alone. Participating schools received hands-on kits, in-person learning sessions, and direct instruction from subject-matter experts, with teaching materials designed to connect space science to sustainability, connectivity, and the Internet of Things. The focus was less on spectacle and more on practical exposure, particularly in regions where access to specialist STEM infrastructure remains limited.
What placed this story firmly in the top twenty was its contrast with much of the year’s EdTech narrative. While many initiatives centered on AI acceleration and automation, this program emphasized foundational skills, early engagement, and long-term workforce development. It underscored a point that resurfaced repeatedly in 2025: global EdTech progress will depend as much on widening access and relevance as it does on advancing technology itself.
16. AI tutor that beat 1.3 million students lands $9.5 million to scale globally
SigIQ.ai entered the 2025 conversation with evidence rather than aspiration. The startup drew global attention after its AI tutor completed India’s UPSC civil service exam in under seven minutes, achieving a score higher than any recorded human candidate in a test taken annually by more than 1.3 million people.
That performance was not presented as a stunt. SigIQ.ai positioned it as a prerequisite. The system was designed to master the exam before attempting to tutor students preparing for it, shifting the debate away from whether AI could support learning and toward what happens when it demonstrably outperforms learners at the very task being taught. The company’s $9.5 million seed round followed shortly after, backing plans to expand beyond India into global test-prep markets.
Unlike AI tools framed around assistance or augmentation, SigIQ.ai challenged assumptions about expertise, assessment, and value. If a system can outperform top candidates under real exam conditions, the question for education systems becomes harder to avoid: how should learning, progress, and success be defined when mastery itself is no longer uniquely human.
15. OpenAI spotlights teacher-led uses of ChatGPT in new classroom collection
OpenAI’s release of Chats for High School Teachers marked a shift away from hypothetical classroom use and toward documented practice. The collection brings together examples written by U.S. high school teachers across subjects including algebra, computer science, visual arts, and social studies, showing how ChatGPT is being used inside real teaching workflows.
Rather than promoting new features, the resource focused on how educators are adapting existing tools to manage planning, assessment, and communication. Teachers described using ChatGPT to draft quizzes, refine assignments, map curriculum, prepare substitute lessons, and review student work against rubrics. The emphasis stayed on function rather than transformation, with AI framed as something that fits around teaching rather than reshaping it.
14. Pulaski County schools roll out AI security technology amid rising safety concerns
Pulaski County Public Schools’ decision to deploy AI-powered security technology across the district reflected a growing reality facing U.S. education systems in 2025.
The rollout of Status Solutions’ Advanced Situational Awareness platform built on an existing partnership that began in 2015. The system provides real-time monitoring and coordinated alerts designed to support faster decision-making during emergencies, while integrating with established school safety procedures rather than replacing them. District leaders emphasized that the technology was intended to strengthen, not overshadow, human judgment and on-the-ground response.
The rollout also reflected a broader shift in how technology was being introduced into schools during 2025. In areas tied to safeguarding and risk, adoption tended to be incremental and tightly framed, with systems positioned as support for existing processes rather than disruptive change.
12. Study finds ChatGPT boosts learning performance but shows limits in complex thinking
A large-scale meta-analysis published in Humanities and Social Sciences Communications added needed clarity to debates around generative AI’s educational value. Drawing on data from 51 experimental and quasi-experimental studies conducted between late 2022 and early 2025, researchers found that ChatGPT had a strong positive effect on student learning performance, alongside more modest gains in perception and higher-order thinking.
The findings were not uniform. Learning gains were strongest in problem-based and skills-driven courses, particularly when ChatGPT was used as a structured tutor rather than an open-ended assistant. Short-term use showed limited impact, while interventions lasting four to eight weeks delivered the most consistent results. In contrast, longer exposure sometimes led to diminishing returns, raising questions about over-reliance and instructional balance.
Equally notable were the gaps. Most of the studies focused on higher education, with limited evidence from primary or early years settings, and weaker outcomes in project-based learning environments. Rather than presenting AI as a universal solution, the analysis reinforced a recurring theme from 2025: effectiveness depends less on the tool itself and more on how deliberately it is embedded within pedagogy, assessment, and course design.
11. Harvey brings generative AI into UK legal education through law school partnerships
Harvey’s expansion into UK legal education marked a notable moment in how generative AI entered professional training during 2025. By partnering with Oxford University Faculty of Law, The University of Law, King’s College London, and BPP University Law School, the legal AI company moved its platform beyond firm-led adoption and into the early stages of lawyer development.
Each institution approached the partnership differently. Oxford focused on research use and experimentation, examining how AI tools could support legal scholarship and pedagogy. King’s College London embedded AI literacy and ethics into its curriculum, while BPP University Law School integrated Harvey into training designed to prepare students for professional practice. The emphasis across all four was exploration rather than standardization, allowing faculty to test where AI added value and where limits remained.
The partnerships reflected a broader shift underway in 2025, as professional disciplines began addressing AI not as an external disruption but as a competency students would need to understand before entering the workforce.
