Pratham x Anthropic Partnership – Using AI to turn Assessment into Opportunity

A new partnership between Pratham and Anthropic is exploring how AI can do something education systems have long struggled with: turn assessment into a moment for learning, not just measurement.

At the centre of the collaboration is the Anytime Testing Machine (ATM), an AI-enabled assessment tool designed to go beyond scores on a page. It generates curriculum-aligned questions, reads handwritten responses, and offers clear, actionable feedback. The goal is straightforward but significant: to help learners understand what they know, where they’re getting stuck, and what to do next.

In many contexts, assessment is still a one-off event. A test is taken, a mark is given, and the process moves on. For learners who are already behind, this can reinforce gaps rather than address them. ATM points to a different approach, where assessment becomes part of the learning process itself, offering guidance that can be acted on immediately.

Crucially, this is not about removing the role of teachers. Pratham’s model has always centred educators and communities, and that remains the case here. AI-generated questions and feedback are reviewed and shaped by teachers, helping ensure they are relevant, appropriate, and useful, particularly for learners working in a second language.

The partnership itself has been deliberately hands-on. Teams from Pratham and Anthropic have worked together over several months to refine how the tool performs in practice, from improving the clarity of questions to making feedback more accessible and easier to apply.

Early pilots in India have already reached thousands of learners, including school students and women returning to complete their secondary education. These are groups for whom flexible, responsive learning tools can have a particularly strong impact.

Looking ahead, the collaboration is expected to expand into AI-supported teaching tools and explore how digital public infrastructure could help scale this kind of approach more widely.

What stands out is not just the technology, but the model behind it. This partnership brings together deep on-the-ground experience in education with cutting-edge AI development, showing what becomes possible when different sectors work towards a shared goal.

If done well, this approach has the potential to make education more responsive, more inclusive, and ultimately more human, ensuring that more children have the chance not just to be assessed, but to progress.