Isabell Baumann, Dominic Harion & Ann Kiefer

Are computers intelligent? Some people might automatically answer ‘yes’ to this question. And they would be even more likely to do so if the word ‘computer’ were replaced with ‘artificial intelligence’ (AI).¹ But let’s first go back a few hundred years and look at the abacus. Is an abacus intelligent? If I first slide two beads across and then three, and I read off the result 5, who did the calculating? The abacus or me? And if I type in 2 + 3 on a calculator and it tells me the result 5, who did the calculating then? A calculator doesn’t know mathematics. Like the abacus, it is only a tool, albeit a more complex one. However, more complex doesn’t necessarily mean intelligent.²

But why is this easier for us to understand for the abacus than for more sophisticated technologies? This becomes very clear using the example of artificial intelligence. Although autonomous AI systems can now be found in almost all areas of life, there is a general lack of knowledge about how exactly these systems work. The resulting urge to blindly trust tech giants makes many people feel insecure and uneasy. The fear of machines taking over is fuelled further by fantasies from science fiction culture.

Developing a critical understanding, as modelled by Alexandre et al. (2021), can serve to reduce these irrational fears. Furthermore, in the 21st century, it is important that young people have a basic knowledge and understanding of AI (Touretzky et al., 2019). In this increasingly complex world, AI – when used ethically – can be a wonderful thing. It holds great potential for individuals and society, as well as for many important fields, such as science, education and industry. However, as long as people don’t understand how it works, AI can only ever be seen and used as a black box, which makes it impossible to innovate (Alexandre et al., 2021).

The module #Pupils vs Machine deals with the skills area of knowledge transfer.³ In particular, this module is about explaining and demonstrating how an AI works. Pupils discover how an AI learns a game and develops a winning strategy. As an example in this module, we use the well-known game Nim. This is played by the pupils in groups. In each group, one pupil assumes the role of the machine and the others play against it. The pupil simulating the machine follows a strict algorithm to come up with a winning strategy at the end.

This is done according to the principle of ‘CS unplugged’ (computer science unplugged). With CS unplugged, pupils work analogue (i.e. without computers) on activities that help them understand a wide range of computer science topics in an interesting way (Nishida et al., 2009). Although the benefits of unplugged activities are not yet widely acknowledged in research (Huang & Looi, 2021), they are particularly valuable for teaching purposes within this key topic on artificial intelligence.

A wide range of learning resources are available for traditional AI topics in particular. The Snap! environment and Google provide a collection of AI experiments for learners.⁴ In addition, Machine Learning for Kids provides online demos with which pupils can learn how to use classification models and then apply them in Scratch.⁵ Many projects however primarily take an application-oriented perspective. However, the underlying concepts of artificial intelligence are difficult to understand in purely application-based situations, which means AI systems remain a black box for learners. Unplugged activities therefore aim to make the underlying concepts of artificial intelligence more accessible. At the same time, they avoid the need for a highly formalised, mathematical representation, which would make it much more difficult for pupils to understand the topic (Lindner et al., 2019).

The module #Pupils vs Machine therefore concentrates less on application areas of AI and more on understanding and demystifying an AI. At the end of the module, the pupils will realise that the machine is not intelligent, just well programmed. To quote Thierry Viéville: ‘If we start believing that computers are intelligent, we will eventually let them rule us. But if we realise that these machines, no matter how complex, are actually just an ocean of calculations, we will have adopted the right attitude’.⁶

^¹ ‘Artificial intelligence (AI) refers to systems designed by humans that, given a complex goal, act in the physical or digital world by perceiving their environment, interpreting the collected structured or unstructured data, reasoning on the knowledge derived from this data and deciding the best action(s) to take (according to pre-defined parameters). AI systems can also be designed to learn to adapt their behaviour by analysing how the environment is affected by their previous actions.’ https://ec.europa.eu/futurium/en/system/files/ged/ai_hleg_definition_of_ai_18_december_1.pdf
² This analogy is taken from the lecture ‘L’intelligence artificielle est-elle si intelligente’ by Thierry Viéville, https://www.youtube.com/watch?v=RH023-y0rJk&t=4s.
³ As indicated in the skills for axis 6 of the subject Digital Sciences.
⁴ https://experiments.withgoogle.com/
⁵ https://machinelearningforkids.co.uk/
⁶ ‘L’intelligence artificielle est-elle si intelligente’ by Thierry Viéville, https://www.youtube.com/watch?v=RH023-y0rJk&t=4s.

Referenzen

Alexandre, Frédéric, Becker, Jade, Comte, Marie-Hélène, Lagarrigue, Aurélie, Liblau, Romain, Romero, Margarida & Viéville, Thierry. (2021). Why, What and How to help each Citizen to Understand Artificial Intelligence? KI – Künstliche Intelligenz, 2, 1610–1987.
Huang, Wendy & Looi, Chee-Kit. (2021). A critical review of literature on „unplugged” pedagogies in K-12 computer science and computational thinking education. Computer Science Education, 31:1, 83–111, https://doi.org/10.1080/08993408.2020.1789411
Lindner, Annabel, Seegerer, Stefan & Romeike, Ralf. (2019). Unplugged Activities in the Context of AI. In: Pozdniakov, S., Dagienė, V. (eds.) Informatics in Schools. New Ideas in School Informatics. ISSEP 2019. Lecture Notes in Computer Science, vol. 11913. Springer, Cham. https://doi.org/10.1007/978-3-030-33759-9_10
Tomohiro, Nishida, Susumu, Kanemune, Yukio, Idosaka, Mitaro, Namiki, Tim, Bell & Yasushi, Kuno. (2009). A CS unplugged design pattern. SIGCSE Bull. 41, 1, 231–235.
Touretzky, David, Gardner-McCune, Christina, Martin, Fred & Seehorn, Deborah. (2019). Envisioning AI for K-12: what should every child know about AI? Proc AAAI Conf Artif Intell, 33(01), 9795–9799.