Preface to this second edition

Welcome!

October 2022 to October 2023

The first edition of this textbook was published in October 2022. Within days, ChatGPT appeared, and we entered into twelve months of AI madness. Every week, new products were launched and improvements to language models and their applications were announced. More importantly, education seemed to suddenly become a benchmark for generative AIs. Teachers and institutions reacted rapidly, acknowledging the arrival of a new tool and incorporating it into the education toolbox – or prohibiting it because of the menace it was felt to convey. There were discussions in the press but also in international organisations; petitions and open letters were signed. The impact on the jobs market was measured, and some companies have already started to replace their workers with AI.

For the authors of this book, all of this resulted in a key question, a challenge, and an opportunity.

The question was the one any author of a technology-linked book is afraid of. Is the book obsolete? This could be the shortest lifespan of any book – just a question of days. The challenge was to aim to include the novelties resulting from the ChatGPT tsunami into a second edition. And the opportunity was to share the book in the best possible moment, when it was probably most needed.

The question – does Generative AI’s importance mean that the rest of AI is now unnecessary?

The question makes sense. ChatGPT has been adopted by many because it is so easy to use. Some generative AI experts of 2023 knew little about AI in 2022. It is therefore tempting to believe that generative AIs are built on thin air and can be understood – if that is the goal – by reading only what has been published in the past year. So, is it still necessary to understand machine learning and the different tools developed by AI techs over the past 70 years?

We believe the answer is “yes”. Even if a spectacular step, generative AI builds on technologies and ideas that have been shared for decades. Understanding data, bias, unsupervised learning, personalisation and ethics is still key to what a teacher should know before using AI in the classroom.

The Challenge

The challenge is to write about a fast-moving technology in a way that would satisfy a teacher who, understandably, wants to work from non-ephemeral knowledge, to build their teaching from concepts and technologies which will be resistant to time. One example is the notion of “hallucination” which has changed so much over the past twelve months and which is going to be crucial to how teachers will adopt generative AIs.

The Opportunity

The opportunity follows the urgency with which all stakeholders are today examining the question of artificial intelligence and education. Whereas in 2020, when the AI4T project was launched, the difficulty was going to be in recruiting enough teachers to learn about AI for the experimental results of the project to be valid. In 2023, this has become a question of the highest priority in all countries.

What is new in this second edition?

We obviously had to take into account the arrival of ChatGPT (and later of alternative AIs). And a whole section (7) is now devoted to understanding the phenomenon; it is starting to propose how a teacher should take advantage of these technologies.

For the more technical aspects, we have chosen   to highlight images over text. There are therefore many new illustrations in this version. We have also added 15 short videos which will – we hope – help to understand important concepts.

The open and multilingual challenge

This is an open textbook which means that a Creative Commons (CC) licence has been used. All images, videos and extra material have been scrutinised in order to be openly shared. This means anyone can take the material, or part thereof, and reuse it as they wish. They can also make modifications. There are different export formats available, and the authors can probably share in any way which ensures that this is a sustainable textbook. It can live on with future versions and new projects.

As is customary, the only obligation is that of citation of the authors of the book or of specific chapters, where relevant.

One particular modification – translation – has been anticipated with as much care as possible. We are already translating the original English version into French, Slovene, Italian and German. Furthermore, new projects are emerging to translate the textbook into other languages. We believe AI can help with the translation process, but human correction is needed.

Please contact us to build a partnership if you want the book translated to your language!

What were we saying one year ago?

Let’s start with what you already know: AI is everywhere and education is not an exception. For some, the future is bright and the coming technologies will help make education available to all; it might even help when there are not enough teachers. It will permit the teacher to spend more time on the ‘noble’ tasks while the machine will take over the ‘boring’ ones, such as grading, organising the classroom.

For others, these AI algorithms represent a huge danger, and the billions of dollars the industry is prepared to invest prove that education is now viewed as a market. But it is not a market.

Somewhere in the middle, between these rather different positions, are researchers, educators and policy makers who are aware of a number of things: artificial intelligence is here to stay and will be in the classroom if it is not there already. And no minister – let alone a teacher – will be able to stop this. So, given this fact, how can the teacher harness the beast and use artificial intelligence for the better? How can the teacher make the AI work for the classroom and not the inverse?

The purpose of this textbook is to support the teacher in doing this. It has been built in the context of the Erasmus+ project AI4T. Teams from Ireland, Luxembourg, Italy, Slovenia and France have worked together to propose learning resources for teachers to be able to learn about AI – specifically AI for education. The learning material and a presentation of the project and its results can be found on AI4T’s webpage (https://www.ai4t.eu/).

Training teachers is an essential task for all ministries involved. The objectives are the following:

1. Making teachers aware of why such training is good. It can’t be an imposed decision; it has to be shared.
2. Introducing AI: from our experience of many conferences and workshops, there are participants who have explored, read and digested the topic. However, the vast majority have not.
3. Explaining how AI works in the classroom. What are the mechanisms? What are the key ideas?
4. Using AI in educative tasks.
5. Analysing what is happening in the field and being active of future changes.

We hope the textbook will be able to help you with most of these questions. We analyse the current situation and link AI with the experience of the teachers. By so doing, we hope to encourage them to remain interested in these questions. Undoubtedly there will be new challenges, mistakes will be made, and there could be strong opposition and controversies. We have sections called ‘AI Speak’ in which we try to explain how and why the algorithms work. Our goal is to inform teachers who can then fully participate in the debates and discussions on education and artificial intelligence. Some reasons for preparing this material can be found in the video prepared by AI4T.

We believe in the following:

• Some AI literacy is necessary. Let’s explain this, as it is often argued that ‘you don’t need to know how engines work in order to drive a car’. This is not entirely true: most of us don’t know how engines work but accept there is science and technology involved. We accept this because in school we received lessons in basic physics and technology. In the same way, we wouldn’t be satisfied with a book telling us not to smoke, based on statistical arguments about the number of smoking-related early deaths. Again, we are able to understand why smoking is harmful because at some point a teacher has explained to us how the respiratory system works, what lungs are, etc. Today, with AI making a huge impact on society, we believe that the same applies – finding out about the effects of AI is insufficient. Teachers need to have an understanding of how it works. The goal is not to make each person a biologist or a physicist – the goal is to make us understand the principles and ideas.
• Teachers are extraordinary learners. They will be critical when something is not explained the right way, and will engage more. They want to understand. This textbook is for people who are prepared to go the extra mile, who will not be satisfied until they do understand.
• Next, AI has to be used in a safe environment – computers or devices will be connected to the web and applications will run on the cloud. A huge security issue exists here, and a teacher needs reassurance that the working environment is safe for all. Computer security is a highly complex question; a teacher will not be able to check the software’s safety specifications. A trusted source will need to do this.
• AI can help, provided it is used in a well-defined and controlled learning environment, for a task which the teacher has identified as important. For obvious economic reasons, the industry will push products on the teachers, ostensibly to help them solve a sometimes-unimportant task. But if it is considered ‘cool’, and is pushed by the seller, it could end up being seen as important. A good teacher should be aware of this. In this textbook, we introduce elements for the teacher to identify such products or situations.
• When preparing this courseware, we did have a serious problem. The idea was to use AI software which we could recommend to the teachers, so that they would rapidly be able to use them in the classroom. Unfortunately, this is not the case: a lot of software is still immature, there are a lot of ethical concerns and in most cases the different ministries and governments have not approved lists of software. We have therefore chosen a different approach: we will be mentioning We had a problem when preparing this courseware. The idea was to use AI software which we could then recommend to teachers, for use in the classroom immediately. Unfortunately, this is not what happened. Much software is still immature and have ethical concerns. In most cases, the various ministries and governments have not approved specific software. Because of this, we will be mentioning software in the textbook. This is because we believe it explains a particular point of AI in education. However, we do not endorse any particular software. It is expected that soon international agencies, such as Unesco, UNICEF or the Council of Europe, will come up with specific software recommendations.

We would like to thank the many contributors who have helped compile this textbook.

First and foremost, we have benefited from reading Wayne Holmes’ works and enjoyed many hours of discussions with him.

Discussions also took place within the AI4T consortium. Workshops were organised in order to get the topics to emerge.

Teachers themselves have been an essential source of information – through seminars and webinars we exchanged ideas with them, and they let us know which were confusing and/or wrong.

Many gave valuable opinions, proofread documents and suggested links and texts. Some added chapters to this work:

• • Manuel Gentile helped us in a number of chapters and showed great skill in making accessible the most obscure aspects of AI;
  • Fabrizio Falchi and Giuseppe Città were great collaborators who have helped us understand a variety of AI questions;
  • Azim Roussanaly, Anne Boyer and Jiajun Pan were kind enough to write the chapter on Learning analytics;
  • Wayne Holmes wrote a chapter on agency. This is an important topic when discussing the ethical implications of AI;
  • Michael Halissy and John Hurley explored the issues of homework and assessment with the arrival of Generative AIs;
  • Bastien Masse is today an expert in mastering  the prompt; he has shared his skills here;
  • Blaž Zupan introduced the Orange software, which his team has been developing, in order to make use of machine learning.

We are also very much indebted to those who coordinated the translated of this textbook into French, Italian, German and Slovene. Our special thanks go to Solenn, Manuel, Daniela and Helena.

La Plaine sur Mer, 26/11/2023

Colin de la Higuera