Morad (Google Research): with AI Quests students learn how artificial intelligence works

Video messages from real researchers working on the scientific projects from which the game takes its inspiration also appear at the end of each course. Each module also includes materials for teachers with preparatory activities and post-game discussions.

In Italia, AI Quests will be integrated within Experience AI, the educational programme developed by Google DeepMind together with the Raspberry Pi Foundation and brought into Italian schools with the Fondazione Mondo Digitale. The project will also be available free of charge to families and educational organisations through the dedicated website.

The initiative comes at a time when the relationship between school and artificial intelligence has become one of the most discussed topics in the world of education. In recent months, the debate has mainly focused on the risks of generative chatbots: copying, cognitive dependency, loss of critical capabilities.

Many schools are still debating whether generative AI is a risk or an opportunity for learning. Is AI Quests also an attempt by Google to shape a new cultural narrative around AI in education?

"Our goal with AI Quests," says Ronit Levavi Morad, Senior Director, Google Research, "is not to impose a single narrative, but to give young students the tools to be the architects and shapers of their own AI-driven future. We want them to move from being 'consumers of AI' to 'builders with AI'. Our researchers are excellent role models in this regard. We believe the best way to prepare them for an AI-driven future is to empower them to be critical, hands-on creators. This game helps shift the debate from "risk versus opportunity" to "how can we be prepared, build and design responsibly?".

I recognise that things are changing at a rapid pace. I think we should listen to all voices as we go through this transformation. We at Google certainly try to do that. That's why my team and I spend a lot of our time meeting with educators, students, policy makers, researchers, entrepreneurs and edtech companies around the world."

 An interesting aspect of AI Quests is that students can make mistakes, test datasets and see how bias or poor quality data affect the results. How important is it for Google that AI literacy includes understanding the limitations and failures of AI, and not just its capabilities?

"It is absolutely essential. In fact, we see it as the most fundamental and rewarding part of AI literacy. By showing students how AI works - including the logic behind the 'black box' and its learning curves - we give them real digital confidence. Understanding why an AI arrives at a specific result helps them hone their problem-solving skills. When students see the character in their game improve because they have improved the data, they experience a powerful "aha!" moment of agency (awareness of their own active role). It turns technology into a transparent and manageable tool that celebrates critical thinking and human creativity.

AI Quests teaches students to collect data, train models and evaluate results within the game. Can you explain what kind of AI pipeline is actually simulated "under the bonnet" and how you have simplified complex concepts such as model training, data bias and validation for 11 to 14 year old students?

"This is a fantastic question because it goes right to the heart of what we built with our partners at the Stanford Accelerator for Learning. A central part of this project was taking complex Google Research and translating it for middle school students without losing the true essence of science.

Under the bonnet, we simulate a true end-to-end machine learning lifecycle: defining a community problem, collecting and cleaning data, training a model and seeing its impact in the real world. But to make it understandable for 11- to 14-year-olds, we had to give room for creativity and anchor everything to tangible, experiential learning:

We use gameplay to make the algorithms intuitive. By testing different data choices, students see how their flood prediction model evolves in River Fair. If they make wrong data choices, their model will perform poorly. They learn through experimentation, refining their approach for better results, mirroring the iterative and creative process of real-world research.

Stanford's research showed us that 'data' as an abstract concept simply does not work with this age group. So, we represented the data within the game using physical SD cards. It's a simple, tangible example, and the kids understood it instantly.

In our medical mission, Twilight Canyon, we introduce deliberate friction. Students build a model to diagnose an eye disease, only to realise that it works perfectly for one group of patients but badly for another. This 'conflict' forces them to actively understand why it is 'unfair' and how to solve it.

As for privacy at Google, this is non-negotiable, so we made sure to integrate it into AI Quests from day one. In the health mission, for example, the very first task for the students is to clean the dataset of names and personal details to protect the privacy of the patients.

At the end of the day, AI is only the means, not the end. We want to show kids that AI is not a black box, but a human-designed tool. By demystifying it, we help them move from passive consumers of technology to architects of AI-driven solutions that can solve real-world problems'.