Instead of viewing artificial intelligence as a tool for task automation, true educational innovation lies in framing technology as a catalyst for collective learning transitions. In this conversation, Olga Kokshagina discusses how we can cultivate critical AI literacy, build human agency and harness AI as an amplifier for collective intelligence in the classroom and beyond.

Could you share your vision regarding AI and the future of education across the broader landscape?

AI and education should work together; with AI acting as a catalyst for enabling and transforming how we learn. I see AI helping us understand complex challenges and systemic change, while also encouraging us to rethink what learning itself can look like. This means thinking about both transitions in learning and learning in transitions: not only changing how we learn, but also developing our capacity to learn and adapt as we navigate a rapidly changing world.

I believe the goal is to use AI to support learning in ways that benefit us as humans. Rather than viewing AI solely as a tool to replace jobs, boost immediate task productivity or automate work, we should explore how it can help us develop new ways of thinking, learning and responding to the complex challenges we face.

You speak about AI as an amplifier of collective intelligence rather than something that replaces human thinking. What does that entail in the educational field, and could you share practical examples from your daily work?

In the workshops that we’ve organised with Anna-Lena Neurohr and Chahab Nastar from Learning Planet Institute, we explored several practical examples involving participants across the educational ecosystem. One that comes to mind is “challenge-based learning”, referring to initiatives in universities where students tackle transdisciplinary issues, such as climate change, sustainable mobility or the future of peace, and AI can help facilitate these conversations.

While students are not expected to solve global challenges single-handedly, this approach builds essential AI literacy and problem-solving capabilities for the future.

Also, we can help to integrate Indigenous knowledge and non-Western epistemologies in AI. Currently, many non-Western epistemologies and Indigenous knowledge systems are underrepresented in AI models. Connecting global networks of Indigenous-led labs allows us to explore how AI design can harmonise Western science with Indigenous knowledge. This shifts technology from extractive paradigms towards regenerative, intercultural ecosystems, empowering learners to incorporate non-human perspectives, such as nature, oceans and wildlife, into environmental design processes.

Is there a specific age group for these workshops? At what age should we introduce AI in education?

The specific workshop reported in Constellations of Hope, included participants ranging from early university students (early 20s) to senior experts. However, AI literacy should involve learners across all age groups. Learning is a lifelong process that spans from early childhood through secondary careers, extending to elderly stakeholders who also need opportunities to develop AI literacy.

For sure, technology exposure must be structured around clear pedagogical intentions. Learners need space to reflect on what they see, ask questions and process disturbing or complex digital content. Also, note that AI literacy does not require immediate screen use. Even with young children, educators can discuss how AI systems work and how data is processed, building a critical foundation before students interact with screens directly.

AI systems can reproduce and amplify bias. How do you work to identify and address these assumptions?

Bias is a central challenge, though it is worth noting that humans are inherently biased as well. AI amplifies human bias because systems are trained on datasets collected by humans, governed by meta-models and processed by specific training algorithms.

That is the reason why, when working in group settings, dealing with bias requires building human agency and critical thinking. Users must avoid accepting AI outputs at face value. Because probabilistic models operate as black boxes, learners must actively question system outputs, discern what is useful and test their own assumptions. AI can even be leveraged to check our own personal biases by asking “what if” questions that challenge initial thinking and reveal unanticipated perspectives.

Could you give an example within education where AI should never be used or allowed to make decisions?

AI should never replace human critical thinking or be used to pass off machine-generated output as original human work. While comparing an AI-generated essay against original human writing can serve as a useful reflective exercise, presenting unreflected AI output as one’s own work is unacceptable.

Educators and researchers face growing challenges with unverified AI integration in peer review and student assessments. Delegating core thinking to AI represents a clear boundary. AI should be used to challenge assumptions and amplify reflection, not to substitute human reasoning.

What practical advice would you give to a teacher looking to integrate AI into their classroom?

I would say to embrace learning by doing, foster critical thinking and encourage collaborative interdisciplinary work.

Related to the first point, you need to know that sometimes students are more advanced than their own teachers in the use of technology. So, learn alongside your students. Create a shared forum where students bring in examples and explore AI applications together. This is closely related to fostering critical thinking. One way to do this is to spend time testing the limits of AI systems, through exercises designed to break their patterns, identify unhelpful or unethical responses and discuss why those failures occur.

And lastly, move away from isolated, individual AI use towards collective learning environments where the entire class reflects on technology together. Invite guest perspectives and use design-based ‘what if’ questions. Prompt students to challenge reality, question assumptions and imagine alternative solutions without feeling pressured to become purely technical coding experts.

This blog is part of a series to introduce Constellations of Hope: Cases and Perspectives of Transformational Learning, a book with a collection of real-life experiences of researchers, educators and practitioners around the world who have explored new learning paradigms.

‘Constellations for Hope’ is a new report to the Club of Rome and is coming out the 4 November at the International Conference of the Club of Rome. Sign up here!

The book is part of No Limits to Hope, a collaboration between The Fifth Element and the WEEC Network.