Edtech has long promised to revolutionise how we learn, only to fall short of expectations. The global digital leap in education – once envisioned as a democratising force – often delivered fragmented experiences, limited personalisation, or tools that failed to reflect how learning actually works.
Yet the pressures on education systems continue to mount. Teacher shortages are worsening, skill demands are evolving rapidly, and the world remains off track to meet the fourth UN Sustainable Development Goal of ensuring inclusive, high-quality education for all.
Against this backdrop, a genuine turning point has arrived. Advances in generative AI, coupled with large-scale learning datasets and clearer pedagogical frameworks, are transforming what educational technology can achieve. For the first time, digital systems, especially AI teaching assistants, are beginning to approximate the responsiveness, feedback, and scaffolding that underpin effective instruction.
Bloom’s well-known “2 Sigma Problem” showed that one-on-one tutoring can lift an average student’s performance to levels typically seen only among the very best. For decades, that kind of impact felt impossible to scale. Today, AI is bringing that level of personalised support within reach for millions.
AI is no longer speculative but a real solution to persistent challenges of access, scale, and personalisation
At Efekta Education, an EF (Education First) company, where I work leading academic development, this shift is not theoretical. We have launched what is among the world’s largest AI-enabled learning pilots, reaching more than four million high-school students across Latin America. In an earlier pilot in Brazil’s state of Paraná, 750,000 students using the platform improved their state English test scores by more than 32% in under two years.
When applied at scale, this represents not just a technological shift, but a structural one. AI teaching assistants are increasingly capable of engaging learners turn-by-turn, diagnosing needs in real time, and adapting tasks to support genuine progress. And the implications extend far beyond tech itself. Governments grappling with stagnant learning outcomes, companies facing skill shortages, and schools contending with overcrowded classrooms all stand to benefit, if these AI assistants are designed and deployed responsibly.
But past missteps offer a clear warning. Edtech has seen cycles of overpromising: products released without proper validation, AI tools that rewarded superficial responses, and high-profile failures where enthusiasm outpaced evidence. These experiences underline the need for discipline and transparency. Without standards for evaluation and clear pedagogical purpose, the field risks repeating mistakes that erode public trust. The promise of AI in education will only be realised if systems enhance, rather than bypass, the human elements that make learning effective.
From the perspective of practitioners working inside this shift, the change is unmistakable. We are seeing how structured learning data, drawn from millions of learners across diverse contexts, can drive AI teaching assistants that genuinely support teaching and deepen learner engagement. These tools extend what is possible in traditional classrooms by offering the practice, personalisation, and feedback that students often lack – especially in large or mixed-ability groups. Crucially, this is not a replacement for teachers but a way to amplify their impact and free them to focus on higher-value guidance and coaching.
Similar AI-driven learning systems built on structured frameworks are now gaining traction globally across the education and corporate markets, and results are promising. In India, companies like SigIQ use AI tutors to help candidates prepare for civil service exams, while the governments of Estonia and Iceland have partnered with providers such as OpenAI and Anthropic to bring AI-powered tutors to every high school student.
OpenAI is also partnering with Khan Academy to scale KhanMigo, an AI-powered tutor currently used by 65,000 American students demonstrating around 20% higher-than-expected learning gains on standard growth assessments. Together, these examples mark a turning point: AI is no longer speculative but a real solution to persistent challenges of access, scale, and personalisation.
This shift is especially significant in fields that underpin economic competitiveness, such as English and STEM. English remains a high-stakes skill: it is the language of innovation, scientific collaboration, and global business. Data from the EF English Proficiency Index (EF EPI) show that countries with stronger English proficiency also tend to have more robust innovation ecosystems and higher GDP per capita.
For employers like Amazon and McDonald’s, which are both using the Efekta platform to upskill their workforces, English is no longer a ‘soft skill’ but a strategic capability in a globally connected economy. Yet the same is true for other essential domains: STEM knowledge, digital literacy, and the critical thinking skills needed to navigate an AI-enabled world. Ensuring broad access to these foundational competencies is becoming central to both national competitiveness and individual opportunity.
The same logic extends beyond large employers. In Rwanda, Efekta partnered with the Ministry of Tourism and the Mastercard Foundation to support English training for hospitality workers, with the majority of participants improving by at least one CEFR level; illustrating how AI-supported learning can expand access to economic opportunity when developed in partnership with public institutions.
What makes this moment distinct is not merely the sophistication of the tools but the alignment of enabling factors: improvements in AI architecture, richer pedagogical datasets, broader acceptance among educators, and the recognition that technology works best when it is integrated into structured teaching environments. This convergence holds the potential to expand access, close skill gaps, and support teachers at a scale that was previously unattainable.
Still, the opportunity is not guaranteed. Without careful governance, AI could exacerbate inequalities, particularly if high-quality tools are available only in well-resourced systems. Policymakers, education leaders, and technology providers all have a role to play in ensuring that AI advances equity rather than deepening divides. That means establishing clear standards for evaluation, safeguarding learner data, investing in teacher training, and ensuring that AI systems reflect linguistic and cultural diversity rather than narrow assumptions.
Like earlier general-purpose technologies, from mobile computing to the internet, generative AI is poised to reshape foundational systems. But its success in education will depend on design choices, ethical commitments, and public accountability. The next phase of edtech will not be defined by novelty, but by whether these tools meaningfully improve learning for the many, not just the few. Done right, AI teaching assistants can close skill gaps, expand opportunity, and drive growth; helping companies, governments, and learners alike find their voice in an AI-enabled global economy.
The inflection point has arrived. The question is no longer whether AI will change education. It already is. The real test is whether we can guide this transformation with integrity and purpose, ensuring that the long-promised benefits finally reach the learners and teachers who need them most.
About the author: Dr. Christopher McCormick is chief academic officer for Efekta Education Group. Christopher is an international education specialist with over 30 years of experience championing innovation in language teaching and learning. As chief academic officer of Efekta Education Group, he oversees the learning design of AI-powered interactive curricula that help to transform education for good. Throughout his career, he has contributed to major programs for schools, companies, and international organisations such as the UN and the Olympics.
Christopher grew up in Texas, then later studied and worked in New Orleans, Paris, Monterrey, Austin, Boston, Shanghai, London, Zug, Lucerne, Colombo, Phnom Penh, Beijing, and Bangkok. He obtained a PhD in Applied Linguistics from the University of Texas at Austin, and he has served in executive roles for EF Education First and as an associate faculty member of Hult Ashridge Executive Education. As a lifelong explorer and the spouse of a diplomat, Christopher shares his passion for intercultural understanding that comes through language learning and cultural discovery.
