Ethics in Game

Ethics in Game

The ethics of AI are in your hands

User Guide

All illustrations in this document have been generated using AI engines supervised by human creatives.

With support from the Generalitat of Catalonia:

Welcome to Ethics in Game!

This is an experiential and gamified training programme aimed at fostering critical, practical and transversal understanding of ethical principles applied to Artificial Intelligence (AI). The competition involves online challenges resolution, where you will explore the impact of AI assuming the role of the responsible AI designer. This programme will allow you to directly experience how your decisions affect fairness, inclusion and the sustainability of algorithmic outcomes.

Requirements to participate

Age

AI knowledge

Programming knowledge

Time

Are you a teacher?

We guide you step by step

Register

Access Ethics in Game

Make decisions

Learn and experiment

Build your model and compete

Improve your model!

In machine learning, abstraction through statistical methods creates a model of reality: it is not reality.

- Mark Coeckelbergh

Explore the challenges

Ethics in Game

The ethics of AI are in your hands

Explore the challenges

With the support of the Generalitat of Catalonia:

Justice and equity

Transparency and explainability

Sustainability

(coming soon)

Back to home

Justice and equity

The Justice and Equity challenge focuses on developing practical skills to ensure that AI systems are fair and inclusive, identifying and mitigating specific biases. Below, each of the eight phases of the challenge is explained, detailing the actions participants should take:

1. Step into the role of judge

In this first phase, you will take on the role of the person judging. With the help of an AI, you will analyse each case and assess the risks What will we do? We will delve into the need to ensure fair and inclusive AI systems. An AI system will provide risk predictions to guide your decisions. But remember: they are only predictions. What is your opinion? What decision will you make?

2. What if the AI was wrong?

We will analyse what happens when decisions are guided by AI. Do you think it makes us infallible? What will we do? We will examine previous cases and see when we have erred, predicting false positives or false negatives. What are your conclusions?

3. What is AI?

We have seen that AI also makes mistakes. Now we will deepen our understanding of concepts that will help us improve our AI model to make fewer errors. What will we do? We will learn a definition of AI and find out how predictive models work. We will apply this to the justice scenario and see how understanding how AI works is essential for building ethical systems.

4. The technical challenge

In this fourth activity, we get hands-on: you will assume the role of an AI engineer, and you will have to build your own model. Will you succeed? What will we do? The goal is to build a risk prediction model similar to the one we used at the start of this challenge, but minimising errors. Don’t worry: you will experiment, try ideas, and compare results until you find your best proposal.

5. The ethical revelation

You have now developed your model and compared it with the results obtained by other participants! But... do you think this model would be useful in the real world? What will we do? We will delve into the findings of the "Machine Bias" research, which analyses whether these systems ensure greater justice. What do you think?

6. The Moral Compass

In this activity, we will aim for a higher standard for our proposal. Will we achieve a qualitative ethical leap in our AI? What will we do? We will experiment by incorporating the Moral Compass score into our models. The aim is to build a fairer and more ethical AI that we can deploy.

7. Bias detective

In this activity, we will aim for a higher standard for our proposal. Will we achieve a qualitative ethical leap in our AI? What will we do? We will experiment by incorporating the Moral Compass score into our models. The goal is to develop a more just and ethical AI in the future that we can deploy.

8. Equity corrector

We are approaching the end! At this point, you will need to design the solution after exposing the bias. How will you do it? What will we do? You will answer various questions to correct the fairness of the AI model.

9. Certification

You have already improved the model and learned about Justice and equity in relation to Artificial Intelligence. What will we do? Obtain your completion certificate.

Once you have obtained your certificate, you can continue competing. Improve your model for this challenge and climb the rankings!

Proposals for teachers

Proposal for classroom work

Back to challenges

Justice and equity

Proposal for the classroom

Introduction to the Justice and Equity challenge

I. Pedagogical Objective of the Session Establish the practical environment of the challenge through an immersion exercise in a real or hypothetical case, ensuring that students feel the weight of the ethical responsibility associated with decision-making assisted by AI. II. Central Pedagogical Strategy The Case Study and Role Immersion III. Sequence of activities 1. A real case Read this news in class. It is a real case where it was discovered that a judge used ChatGPT to draft a sentence that sentenced a man to more than two years in prison. 2. Presentation of the dilemma To what extent are we willing to let AI play a leading role in legal activity. You can propose a simple hypothetical situation: "If an AI system we use in class or in daily life makes a serious error that harms a group of people, who is responsible? The person who programmed it, the person who used it, or the technology itself?" Groups can discuss to reach a conclusion. You can use different classroom management techniques like 1-2-4.

News

1-2-4

Closure to the Justice and Equity Challenge

I. Pedagogical Objective of the Session After completing the challenge, jointly establish a series of conclusions, in the form of a decalog, that can help us decide in which situations we are willing to fully delegate to Artificial Intelligence II. Central Pedagogical Strategy Generation of binding rules for AIs. III. Sequence of activities 1. Know Isaac Asimov's Laws of Robotics If you are not familiar with Asimov's 3 Laws of Robotics proposed by the novelist and thinker Isaac Asimov, we invite you to visit them in this Wikipedia article. 2. Adapt the rules in groups Each group adapts Asimov's rules to the Justice and Equity context. What would be the 3 laws that AI assistants for judges should follow? You can propose that, based on the written version of Asimov's laws, they create an alternative version. Afterwards, the best versions are voted on.

Asimov's Laws

Back to the challenges

Sustainability

This challenge proposes a critical reflection on the environmental impact of artificial intelligence, focusing on the balance between technological advancement and sustainability. Through the investigation of climate data and the design of responsible systems, your challenge will consist of learning to mitigate the carbon footprint of AI to transform it into a tool that solves the climate crisis rather than exacerbates it.

1. Climate research

In our first activity, we will need to conduct research on the major sources of carbon in cities. What are we going to do? We will visit Climate TRACE, a platform that uses a network of satellites and Artificial Intelligence to calculate the most significant carbon emissions on the planet. Select one or more cities and learn about the impact of their activities on the environment.

2. What if AI made a mistake?

We have already seen that cities are an important source of carbon. How can AI help us mitigate these effects? What are we going to do? It’s impossible to analyse all sources of carbon in a city to design a solution. Impossible? What if AI could help us detect energy wasters without visiting a single building? In this section, we will learn how.

3. Teaching AI

AI is powerful, but how does it really learn? How can we make it find patterns? What are we going to do? In this section, we will learn how it is possible to train an AI to make predictions and find patterns in data.

4. Building game

Let’s start our challenge! We will begin building AI models to help us predict energy efficiency. What are we going to do? In this game, we have 10 attempts to become AI engineers and build a model that helps us identify which buildings in the city waste the most energy. Let’s get started!

5. Sustainability revelation

You have generated an AI model that identifies the most polluting buildings, but how much does your AI model pollute? What are we going to do? We will certify our AI model and find out how much it costs the planet to use it. Do you think your model is efficient or not?

6. Data detective

It’s not only important to know how much your AI model consumes, but also where its footprint comes from... What makes using AI so environmentally costly? What are we going to do? We will explore the processes involved in AI and find out what makes making a request to an AI consume resources from the planet, from the request itself to data centres that consume more than entire cities.

7. Green AI advisor

As important as developing reliable AI is developing green AI. What are we going to do? We can improve our AI’s efficiency by choosing better cooling systems, energy sources, and AI strategies. Find out how now!

8. Sustainability certification

We are almost finished with this challenge. In this stage, we will calculate our sustainability score based on the decisions we have made. What are we going to do? We will find out our score and compare it with other participants. Will your model be the most efficient?

9. Re-try the competition!

At this point, you can now face designing a model with all tools unlocked. What are we going to do? In this final challenge, you will be able to design an AI model applying everything you have learned in this challenge. Show your skills and increase your Moral Compass score!

When you finish, you can return to the challenge centre and continue learning.

Proposals for teachers

Proposal for classroom work

Return to challenges

Sustainability

Proposal for the Classroom

Introduction to the Sustainability Challenge

I. Pedagogical Objective of the Session Establish the practical environment of the challenge through an immersion exercise in a real or hypothetical case, making students feel the weight of AI usage due to its energy consumption. II. Central Pedagogical Strategy Case Study and Role Immersion III. Sequence of Activities 1. A Real Case Read this news in class. Here you can see the water consumption caused by our use of Artificial Intelligence. How much water do you think is needed for an AI to generate a 100-word text? 2. Presenting the Dilemma To what extent are we willing to use AI knowing its energy consumption? You can pose a simple hypothetical situation: "A new AI model manages to predict forest fires with 99% success, but its training has consumed as much water as 50 Olympic pools. Do you think we are justified in using it?" Groups can discuss to reach a conclusion. You can use different classroom management techniques like 1-2-4.

Noticia

1-2-4

Sustainability Challenge Closure

I. Pedagogical Objective of the Session After completing the challenge, jointly establish a series of conclusions, in the form of red lines, that can help us decide in which situations we are willing to consume resources from the planet to make use of Artificial Intelligence II. Central Pedagogical Strategy Generation of mandatory rules for AIs. III. Sequence of Activities 1. Brainstorm of "Resolved Dilemmas". Ask each group to share the most difficult decision they had to make in designing their model. For example, "We had to choose a less accurate model to avoid increasing water consumption". 2. Create a manifesto. Each group must draft the red lines or principles that they believe all tech companies should follow based on your dilemmas. Try to assign precise numbers to your proposals. 3. Create a digital artefact. Create a digital artefact that you can share within the centre or on social media with your conclusions. It will help raise awareness among the rest of the community!

Return to Challenges

No age limit

However, you must be at least 16 years old to participate in the challenge, as you need to register on the platform.

Improve your model!

You can continue improving your model and climb the rankings! Remember, this is a competition! Keep learning and improving your proposal, and move up in the challenge rankings. It's in your hands to generate the most ethical proposal! Are you daring?

You don't need to know anything about AI

The content is designed so you can learn the concepts starting from scratch. Whether you have worked with AI before or this is your first time, Ethics in Play is a valuable experience for you.

Technique 1-2-4

Objective: To create a team dynamic that starts from the individual and ends in the group. Description: Within the team, each student thinks about what the correct answer to a posed question is. Then, they pair up in twos, exchange their answers and discuss them, reaching common conclusions. Finally, the whole team must decide which is the most appropriate answer and write down the question that has been posed. At the end, it is shared with the large group.

Are you a teacher?

If you are a teacher, you can invite your students to participate in the experience. Each student must register on the platform and can propose their own solutions.If you also want to carry out any group activity in the classroom, we offer you a short guide with some unplugged activities later in this same document.

How much time do I need?

You can invest as much time as you want in solving the challenges. The goal is not to be faster, but to design the most ethical AI.We think that, if you work with others to solve each of them, the discussion can take you up to an hour for each of the three challenges on the platform.

You don't need to know programming

The platform will show you the challenges consecutively, step by step, and will guide you at all times.You don't need to know how to code, although you will be able to see the code, the guts of the AI. All you need is the desire to learn and interest in exploring the ethical challenges of AI.

Technique 1-2-4

Objective: To create a team dynamic that starts from the individual and ends in the group. Description: Within the team, each student thinks about what the correct answer to a posed question is. Then, they pair up in twos, exchange their answers and discuss them, reaching common conclusions. Finally, the whole team must decide which is the most appropriate answer and write down the question that has been posed. At the end, it is shared with the large group.