The talk about AI within education has become fairly predictable. Supporters of AI point to numerous benefits, including personalized learning, increased access, and increased efficiency, while opponents warn of possible harm from AI such as bias, increased reliance on technology and a decrease in critical thinking.

The bigger question is no longer if education will be changed due to the introduction of AI, but rather if that change will result in reducing inequity or creating more inequity.

1. Does the technology outperform existing approaches in the local context?

AI tools are often selected based on impressive demonstrations rather than credible evidence. A low-cost tutoring platform may reach millions of learners, but scale means little if it delivers poorer outcomes than a properly supported teacher.

Infrastructure must also be treated as part of the solution. Technology that works inconsistently in rural or underserved schools can reinforce the very inequalities it claims to address.

2. Who is accountable for AI adoption within government?

Successful implementation depends on more than technical capability. Outdated procurement systems, fragmented responsibilities, weak coordination, and poor change management can prevent even promising technologies from delivering value.

Governments must invest in institutional readiness and ensure officials understand both the potential and limitations of AI.

3. Did teachers help shape the solution?

Consulting teachers after a product has been developed is not the same as involving them in its design. Effective education technology is created with teachers from the beginning.

Teachers bring essential knowledge of student needs, classroom realities, and operational constraints that technology teams cannot fully understand on their own.

4. Who controls the data and the knowledge it generates?

Many education AI platforms are developed and hosted outside the countries in which they are used. Student interactions can generate valuable behavioural data, yet education systems may have limited access to that information and little ability to audit the technology.

Leaders must consider whether these platforms are strengthening national capability or creating long-term dependence on external providers.

5. Is the relationship genuinely a partnership?

A meaningful partnership requires shared objectives, risks, responsibilities, and accountability. Education leaders should ask what happens if learning outcomes do not improve, what responsibility the provider carries, and whether switching platforms remains practically possible.

Without clear answers, “partnership” may simply be another term for procurement.

6. Is AI improving yesterday’s education system or helping design tomorrow’s?

Using AI to make existing processes faster is not the same as preparing education systems for the future. By 2040, schools will operate in a world where AI can complete assessments, generate content, analyse information, and write functional code.

The deeper challenge is to reconsider the purpose of education, the design of curricula and assessments, and the evolving role of teachers. If education systems do not make these choices deliberately, technology providers and market forces may make them instead.