Reviews and Resources
There has been an explosion of work in the area of AI and Ethics over the last ten years. Below are a few suggestions of texts and resources that those new to the area might find useful.
Atlas of AI: Power, Politics and the Planetary Costs of Artificial Intelligence, Kate Crawford, 2022
An excellent, readable starting-point for anyone particularly interested in the ethical challenges that AI creates. Crawford examines AI as a potential "registry of power", nothing that "Contests around technology are always linked to larger struggles of economic mobility, political manoeuvring and community building."
​
The 'atlas' here is a well-chosen title: there is a continual attempt to focus on AI in its wider global application, and to highlight the dangers of too great a Western, English-language-focused model. This is not a clinical overview, but a passionate, witty and personal investigation by a social activist.
​
Our one criticism is that Crawford is too quick to accept the premise of the 'blackbox problem' (see the Oxford Handbook of Ethics of AI, Introduction for a solution).
AI Ethics (MIT Press Essential Knowledge Series), Mark Coeckelbergh, 2020
Pithy and gently magisterial, this book is a clear and calm overview of the main areas of concern. It is measured in its outlook, with Coeckelbergh refusing to be drawn by extreme views of either side of the debates. He is honest where the areas of uncertainty lie, which, given the speed at which AI has been developing more recently, is everywhere.
​
Coeckelbergh is a Professor of Philosophy of Media and Technology at the University of Vienna, as well as a member of the EC's High Level Working Group on AI. He writes of the potential reasons for optimism with AI, but is direct in addressing its current limitations and the possibilities for ethical regulation. Coeckelbergh is part of the call for the 'right to explanation' for the blackbox problem, but is also honest about the impossibility of using unbiased data.
​
"... Many actors in the AI field... still think of ethics as a constraint, as something negative. This idea is not totally misguided: often ethics has to constrain, has to limit, has to say that something is unacceptable. And if we take AI ethics seriously and implement its recommendations, we might face some trade-offs, in particular in the short-term. Ethics may have a cost: in terms of money, time, and energy. However, by reducing risks, ethics and responsible innovation support the longterm sustainable development of businesses and of society." (p.174)
Artificial Intelligence, A Guide for Thinking Humans, Melanie Mitchell, 2020
Mitchell is an academic working in the field of AI (she is an External Professor at the Santa Fe Institute), but adopts a rather more accessible tone in this book. Her first few chapters are full of student enthusiasm and warnings of terror from seasoned AI experts, before she settles into detailed explanations of how AI works. For someone who wants to understand the technology and maths behind AI, this might be the book for you.
​
After taking us through a brief history of AI, and an explanation of its basic underpinnings, Mitchell moves into the ethical problems surrounding AI. She is fair and distanced in her assessments, but maintains a note of caution throughout. You suspect that she has had rather enough of breathy tech-bros pontificating on utopian AI futures.
​
"Should we embrace the abilities of AI systems, which can improve our lives and even help save lives, and allow these systems to be employed ever more extensively? Or should we be more cautious, given current AI's unpredictable errors, susceptibility to bias, vulnerability to hacking and lack of transparency in decision-making?" (p.145)
The Road to Conscious Machines: The Story of AI, Michael Wooldridge, 2020
Wooldridge is Head of the Department of Computer Science at the University of Oxford, so is well-placed to chart the development of AI, knowing many of the individuals involved.
​
The book is very much a history and takes you through the key developments on the journey chronologically. As a result, you may wish to move through to halfway through the book when he explains the rise of deep learning. A particular strength of the work, we felt, was the way he maps the relationship between neuroscience and AI - something that perhaps has rather been diminished in other text as knowledge-based AI has fallen out of favour.
Oxford's Institute for Ethics in AI
Home | Ethics in AI (ox.ac.uk)
​​
Not a review, but a recommendation. Sign up to their newsletter for links to the blog​. This has thought-provoking essays on a regular basis and demonstrates useful links to other similar intiatives.
FullyHuman
​
FullyHuman is the PSHE Association's research arm, posting cross-disciplinary ideas, articles, films and pocasts to examine the role of technology on children and education. Even if you are not directly involved in education, it provides a punchy insight into the challenges the future generation faces. While not one for expert, business-orientated commentary on AI, its pieces will promote discussion. Just ignore the zeitgeisty coffee-house graphic design.