Learn to Machine Learn — 7 Free Games that Teach Artificial Intelligence

Jasper Schellekens
6 min readApr 17, 2020

--

The famous Chinese philosopher, Confucius once said: ‘I hear and I forget. I see and I remember. I do and I understand.’

This list of games to learn Artificial Intelligence while playing is perhaps a little late for him (by about 2500 years), but perhaps still in time for you.

Free Educational Games for Lockdown Learning

Last month we shared our top free educational games for lockdown learning and now, as promised, we have another list of educational games. This time we’ll provide a list of games that teach artificial intelligence, a critical skill for this generation as algorithms are everywhere. We have developed our own free prototype as part of the LearnML project, but there are a lot of free resources out there already.

Artbot by the Institute of Digital Games

ArtBotis a game for players of all ages which teaches the basics of Artificial Intelligence. Your quest is to find and retrieve stolen art objects. You train your AI helper to recognise and locate the objects hidden in a maze of dungeons, and see how supervised and reinforcement learning works. The game introduces players to core principles and concepts of Artificial Intelligence.

By setting the game in the context of cultural heritage our aim was to address the application of AI systems in multiple different areas, beyond computing and programming, such as archaeology, art, and transportation.

AI Games for All Ages

Minecraft Hour of Code. Source: minecraft.gamepedia.com

Minecraft Educational Edition : Hour of Code — Save your world from a fire by learning the basics of coding AI! Some installation required. The Demo Edition is free.

That sounds great, but what does it teach?

This fun game teaches automating processes via AI. It also teaches the importance of sensors and gathering data for artificial intelligence agents.

Semi-Conductor — Always wanted to be a composer? Now is your chance! Conduct an AI orchestra with your body. It is browser-based and no installation required, but it does require a camera. I tried this with my 2-year old and with my partner, both of them enjoyed it. Bonus points because you get some physical movement.

Semi-conductor in action

Shadow Art — Learn how to make shadow puppets with the help of AI! Another browser-based game that requires a camera. Puppets are pretty challenging to make, so it may be difficult for very young children.

That sounds great, but what does it teach?

It shows how computer vision and artificial intelligence can interact in a fun way. Players will understand that the camera tracks specific parts of the body and assigns these to control the music or the shadows. These are mainly fun tools, but provide a starting point into exploring AI, computer vision, and affective computing.

Thing Translator — Test the limits of computer vision while learning a new language. It is basically a translator for pictures. Requires a camera.

That sounds great, but what does it teach?

Primarily it is a good tool to explore the limits of artificial intelligence and computer vision. When I tried it out I tested the level of detail it could identify. A super-close up of my headphones was identified as “gadget” and further away it was correctly identified as “headphones”. Insight into identifying features used by the algorithm. And also you learn what they’re called in a different language. That’s right, I have my very own “auriculares”.

Tynker — A complete platform focused on using games to learn about coding. Tynker has a number of levels, so probably everyone from 4+ can find something they can work with. All the way from the beginning of coding (with games similar to Turtle Logic parents may be familiar with) all the way to writing actual code in Python. Parents can set up accounts for their children and monitor progress. It is a learning platform and there is a lot of material and a lot of options which may seem daunting at first. While Tynker is normally a paid service with some parts free, during Covid-19 school closures much of the premium content is available for free.

That sounds great, but what does it teach?

While none of the units available specifically cover AI, they do teach coding fundamentals. Furthermore there is even a course to teach Python, currently one of the prefered languages for AI researchers and practitioners. Modules cover IoT and robot sensors central to some AI applications.

Machine Learning for Kids — This is a free tool that trains a learning algorithm to recognize text, words, images, or numbers/sets. The fun part is that children can use the trained algorithms on coding platforms Scratch and App Inventor and build games.

That sounds great, but what does it teach?

It teaches children the relationship between the input and output of algorithms. It also tests some of the knowledge with quizzes while the model is being trained. It also forces children to think creatively about how they can apply this in practice.

Game-based AI Learning — Machine Learning Games for Teens and Adults

Evolution Simulator — Start from generation one and simulate the evolution of a walking robot. The browser based evolution simulator shows how a structure evolves over time. The interface isn’t particularly intuitive initially, so you’ll need to make time to watch the Youtube Tutorial, but after that it is very interesting to view the evolution of the structure over time. It is pretty complicated material and would be recommended for older children and teens.

That sounds great, but what does it teach?

Funnily enough, it teaches both biology and machine learning. Artificial intelligence is often inspired by biology and evolutionary algorithms are one of the ways that machines can learn something. This provides a good overview of how an evolutionary algorithm will develop over time.

Bonus Game: More than Just Machine Learning — Ethics of Artificial Intelligence

Moral Machine — MIT has developed this thought experiment about self-driving cars. It provides scenarios where the self-driving car fails and you need to provide it with what you think is the most ethical decision. A modern take on the classic trolley problem to examine AI ethics. You can also choose to help with their research at the end.

What should robot drivers do?

The Institute of Digital Games is ranked as one of the top 25 graduate schools in game design by the Princeton Review (#16 in 2020). It is a multidisciplinary centre studying games, focused on game analysis and game technology. The Institute of Digital Games is performing cutting edge research at the crossroads of AI, games, and humanities pushing the boundaries of the field with innovate work awarded at numerous conferences and competitions, collaborating actively with industry giants such as Ubisoft. In addition to academic research supported by European and national funding.

Special thanks to Dr. Iro Voulgari, post-doctoral researcher at the Institute of Digital Games for her input. She is an expert in game-based learning. Her research focused on The Value of Computer Games in the Classroom: Qualities, Applications, and Evaluation of Effective Learning Tasks. She also hold a PhD from the Department of Educational Sciences and Early Childhood Education, University of Patras, Greece with her research focusing on Massively Multiplayer Online Games & Learning.

Re-published from the Institute of Digital Game website with permission. Original can be read here.

Edit 24/03/2021 — Updated to include Artbot game made at the Institute.

--

--

Jasper Schellekens

Jasper Schellekens is part of the top-ranked Institute of Digital Games at University of Malta. Leading innovative research from Game AI to Game Philosophy.