Life 3.0 - Being Human in the age of Artificial Intelligence by Max Tegmark
All blog posts are my personal musings and opinions. They are not intended as legal advice.
At a glance
Read this book. I cannot recommend it highly enough. I would say this a is a must-read for absolutely everyone as it will enable intelligent conversation and debate on some of the most profound issues that humanity will have to grapple with in the face of Artificial Intelligence.
For lawyers, I recommend it even more. For me, the GDPR and the recent European Commission’s guidelines for trustworthy AI (currently in draft format as at Jan 2019) were suddenly more clear and far more nuanced than I had initially thought. A good understanding of the ethical issues should underpin all legal analysis when it comes to AI, and Life 3.0 gives you a headstart by providing a wealth of information on the current state of research and the critical issues that prominent AI scientists are currently grappling with.
Top tip: my favourite way to get through lots of books is to listen to them. Life 3.0 is available on Audible. If you don’t have an account then you can try it here: https://amzn.to/2FpYFit
The book starts with a bang and the prelude sets the scene for many of the interesting discussions that follow. The following chapter, ‘Matter Turns Intelligent’ nearly lost me, and I won’t pretend that there was not a copious amount of googling to make sense of what Tegmark clearly thought was a very elementary introduction to the topic. Easy for a Cosmologist to say.
If you persevere and make it through the mind-boggling section, you will be well rewarded. What follows is an incredibly insightful and eye-opening account of the state of AI research and the need for AI safety to be taken seriously.
Lawyers will no doubt find the portion of Chapter 3 dedicated to an analysis of the law particularly interesting. The section is not very long and covers one example of practical application (RoboJudges) and some legal controversies regarding the evolution of regulation (i.e., how laws should evolve to account for technological advances). Although the section is not detailed, it is certainly not the only chapter relevant for lawyers. If you consider that the general theme of the book is essentially an argument (if rather lengthy) for the need to think seriously about the potential implications of AI and to take steps now to safeguard the future, then the whole book should be of interest to the legal profession as a call (and potential precursor) to regulation.
As a fellow Swede, I may be biased, but I found Tegmark’s humour and style of writing both witty and compelling. Making light of certain controversies enables the reader to get a better overview of the issues where broad agreement is found across the scientific community, as opposed to the topics of disagreement so often highlighted in the press. The need for urgent and comprehensive policy agreements on AI safety being the most relevant such example.
Some big questions
Life 3.0 tackles a large number of fundamental questions for AI development. Here is a selection:
What might an intelligence explosion look like? Would it be a slow takeoff with multipolar scenarios or fast takeoff with a unipolar outcome? Which is preferred?
As a super-intelligent AI develops a builds an accurate model of the world and of itself, how would it relate to being controlled and confined by intellectually inferior humans with goals it may understand but that we have no basis for thinking it would share.
On the topic of goals, would the best course of action be to ensure that any AI shares our goals? How do goals work (in terms of physics, biology, engineering, and psychology)? What are the ethical considerations of choosing goals and what should the ultimate goals be?
Consciousness is a whole chapter in itself and presents some genuinely mind-bending questions. Is machine consciousness something that we should strive towards, care about or seek to avoid? Why would it matter? What is the current state of science and what does it tell us about the potential outcomes.
If you are interested in these questions, then Life 3.0 is definitely for you. You won’t find answers (many of these are still unknown), but you will understand the issues and the more intellectually curious people that can contribute to the debate the sooner we can start forming some conclusions on what we would like the future to look like.
Although Tegmark calls for urgent attention to AI safety and the need to take steps now to safeguard the future of humanity, there is also something optimistic about the book that left me feeling inspired and challenged rather than convinced of our inevitable demise. Not everyone shares this view.
In preparing this review, I had a look at what others have written on the topic. For example, Yuval Noah Harari (another favourite author of mine) published a review for the Guardian in 2017 where he observed that:
“Though Tegmark is probably correct in taking things to this cosmic level, I fear that many, if not most, of his prospective readers will not follow him there. Our political systems, and indeed our individual minds, are just not built to think on such a scale. Current political mechanisms barely manage to make decisions on the scale of decades – how can they make decisions on the scale of millennia? Who has time to worry about AI taking over the planet when you have to deal with Donald Trump and Brexit?
In the case of the AI revolution, as so often before in human history, we will probably make the most profound decisions on the basis of myopic short-term considerations. The future of life on Earth will be decided by small-time politicians spreading fears about terrorist threats, by shareholders worried about quarterly revenues and by marketing experts trying to maximise customer experience.”
Harari is not wrong in making this observation but, importantly, the problem is not with the technology (the future of AI is not deterministic and can be shaped by action) but with our current political and economic models. Here, I share Harari’s concern but I also feel that the recent guidelines by the European Commission indicate that there is a political will to at least start to address some of these big questions. As the discussion is inherently regulatory in nature, an informed legal community can also do much to advance these issues.
Life 3.0 is a book that I will keep on the bookshelf and read again in the future, I hope that the next time I pick it up we will have made progress, but I no longer feel that progress is solely the responsibility of others. In that spirit, I launched this blog while reading Life 3.0 and I hope to bring awareness of these issues to the legal community.
The next time I pick up this book, I’m going to ask myself what I have done to ensure progress rather than just hoping that others will have done so.
If you want to join the discussion, Tegmark (and others) have launched The Future of Life Institute a volunteer-run research and outreach organisation that works to mitigate existential risks facing humanity.