The TAILOR Handbook of Trustworthy AI#

An encyclopedia of the major scientific and technical terms related to Trustworthy Artificial Intelligence

This book (to be consolidated in the second phase of the project) represents the first period deliverable of the TAILOR project, providing an encyclopedia of the major terms related to trustworthiness.

About the TAILOR Handbook#

This is a working document for the Version 1 of the D3.3 Handbook on Trustworthy AI, the TAILOR WP3 Handbook on Trustworthy AI. This is a TAILOR project deliverable with two versions: Version 1 (M22) and Version 2 (M46).

The Handbook on Trustworthy AI assumes an encyclopedia-like structure and is presented in the form of a publically accessible WIKI. To do so, the Jupiter Book framework has been used. In the long term, the Handbook is meant to become a point of reference for resources (key concepts, tools, documentation, tutorials, teaching material, etc.) related to Trustworthy AI.

Here, you can find information about the topics of each of the task of WP3, summarizing one of the aspects of Trustworthy AI; the order of the chapters that are in this Handbook simply reflects the order of the task in the workpackage. In particolar, in this Encyclopedia you can find definition related to:

  • Explainable AI. In this part of the Handbook, we will provide an overview of the main properties that an explanation should have and of the several methods to provide multimodal explanations; moreover, our focus will be also on overcome the need of expaining opaque model and, instead, move towards the use of transparent models.

  • Safety and Robustness. In this section of the Encyclopedia, we will analyze the challenges in developing AI systems that are safe, reliable, and robust; we will also provide a way to evaluate this aspects in practice, and we will promote the dynamic evaluation in managing risk during the normal use of AI systems.

  • Fairness, Equity, and Justice by Design. In this chapter, we will start recalling what the grounds of discrimination are, how we can define a bias or segregation; then, we will make a step in defining what fair machine learning could be, and what are the metrics we can adopt to measure (un)fairness.

  • Accountability and Reproducibility. Here, we analyze the two souls of this topic, i.e., the two interrelated concepts of Accountability and Reproducibility: the former term is more related to responsability, blameworthiness, liability, and prevent misuse, where the latter term is more related to measures, quality standards, and procedures to model the development of learning methods for AI.

  • Respect for Privacy. This section will provide an overview of the main attack that can put at risk individual privacy, we will explain the difference between pseudonymization and actual anonymization, and we will describe the main family of privacy models.

  • Sustainability. The last chapter of the Handbook is focus of one of the newest challenge that our society is facing, in particular, our focus is to provide solutions for optimizing both the resources used in AI systems and the computation itself.

Finally, we report a final chapter, where you can find an Index, which lists all entries in alphabetical order; in each term you can find a reference to a short definition of the entry and where it is used within the Handbook, with the link to go more in dept with the definition. Potential synonyms have their own entries in this index.

Executive Summary#

WP3 management and task leaders decided to transform one of the deliverable of the TAILOR EU project in a encyclopedia-like document and present it in the form of a publically accessible WIKI. To do so, the Jupiter Book framework has been used, which is an open source project, supported by an open community of contributors, many of whom come from the Executable Books Community and the Jupyter community.

The main goal of the Handbook of Trustworthy AI is to provide to non experts, especially researchers and students, an overview of the problem related to the developing of ethical and trustworty AI systems. In particular, we want to provide an overview of the main dimensions of trustworthiness, starting with a understandable explaination of the dimension itsleves, and then presenting the characterization of the problems (staring with a brief summary and the explaination of the importance of the dimension, presenting a taxonomy and some guidelines, if they are available and consolidated), summarizing what are the major challenges and solutions in the field, as well as what are the lastest research developments.

However, each entry will be correlated with a bibliography, allowing the reader to go more in depth with a specific topic.

All the entries have a list of authors that have directly contributed to the writing (some of them are already external to the TAILOR consortium), while the complete list of contributors can be found here.



Giovanni Comandé, editor. Elgar Encyclopedia of Law and Data Science. Edward Elgar Publishing, 2022. ISBN 978-1.83910-458-9. URL:


Juan Ramón Rabuñal Dopico, Julian Dorado, and Alejandro Pazos, editors. Encyclopedia of Artificial Intelligence (3 Volumes). IGI Global, 2008. ISBN 9781599048499. doi:10.4018/978-1-59904-849-9.


Claude Sammut and Geoffrey I. Webb, editors. Encyclopedia of Machine Learning and Data Mining. Springer, 2017. ISBN 978-1-4899-7685-7. URL:, doi:10.1007/978-1-4899-7687-1.


Aris Gkoulalas-Divanis and Claudio Bettini, editors. Handbook of Mobile Data Privacy. Springer, 2018. ISBN 978-3-319-98160-4. URL:, doi:10.1007/978-3-319-98161-1.