AI: A Tale of Two Futures.

“AlphaZero crushes chess!”  scream the headlines[1] as the AlphaZero algorithm developed by Google and DeepMind took just four hours of playing against itself (with no human help) to defeat the reigning World Computer Champion Stockfish by 28 wins to 0 in a 100-game match.  Only 4 hours to synthesise the chess knowledge of one and a half millennium of human creativity! This followed the announcement just weeks earlier that their program AlphaGoZero had, starting from scratch, with no human inputs at all, comprehensively beaten the previous version AlphaGo which in turn had spectacularly beaten one of the world’s top Go players Lee Seedol 4-1 in a match in Seoul in 2016.

Interest in AI has reached fever pitch in the popular imagination – its opportunities and its threats. The time is ripe for books on AI and what it holds for our future such as the two recent books [1,2]. Both books agree on the boundless possibilities of AI but apart from this, they could not be more different.

First, they reflect the personalities of their authors – on one side is Max Tegmark, who is Professor of Physics at MIT, but looks more like a rock star with his cool leather jacket and communicates with high flying folksy flamboyance. His book is hardly about AI at all, save one chapter where he quickly compresses how “matter turns intelligent”. For the rest, it ranges over vast magnitudes of space and time as befits a cosmologist: ten thousand years in one chapter, and if that isn’t enough, a billion years and “cosmic settlements” and “cosmic hierarchies” in the next!

On the other side is Toby Walsh, the typical staid university professor, no leather jacket and feet firmly on the ground. Walsh is a computer scientist who has worked in AI for several decades. Part 1 in his book gives a survey of how AI developed from the seminal paper of Alan Turing, through the early days of GOFAI (“Good Old Fashioned AI”) to modern statistical machine learning and Deep Learning. This is a fairly accurate evolution of the discipline and the different “tribes” in it, a compressed version of the account in Pedro Domingo’s Master Algorithm [3] from a few years back. Part 2 is about the present of AI and Walsh gives a panoramic survey of the state of the art in areas such as automated reasoning, robotics, computer vision, natural language processing, and games. He discusses the AlphaGo  program – he is somewhat inaccurate in not mentioning that a key ingredient in it is reinforcement learning. He voices some skepticism about how general the technique is, conjecturing that it “would take a significant human effort to get AlphaGo to play even a game like chess well” – the perils of forecasting! Just a year later AlphaZero used the same principles to crush chess with no human input! However, Walsh is right to point out that chess and Go are limited perfect information games and there is still a long way to go towards more complex interaction games such as Poker, let alone real world problems such as autonomous driving and robotics.

As to the future of AI, both books agree that in principle, superintelligence is possible: that machines could, in principle, become more intelligent than humans, as indeed Turing contemplated in his original paper from 1950. For Tegmark, there are no physical laws that mandate against it and for Walsh, there are no computational principles that preclude it.

When is this likely to happen? Here the two books could not be more different. For Walsh, this is today only a very distant possibility, at least a few hundred years away if not more. Tegmark, on the other hand, seems to suggest that this is just around the corner, and moreover that a sizeable number of AI researchers also think so. A poll conducted by Müller and Bostrom from the Future of Humanity Institute (FHI) is often cited as proof of this but Walsh takes a closer look and highlights some glaring deficiencies, in particular that the poll completely missed the most relevant community of researchers. He points out that a subsequent poll that did target the right group of researchers – namely those who had “made significant and sustained contribution to the field over several decades” came to very different conclusions. Another even more recent poll by the FHI group, this time targeting AI experts [4] also came to somewhat more nuanced conclusions.

So, as AI advances rapidly, what are the future risks? Here again, they agree on a few things. Both are seized of the dangers of autonomous weapons and have devoted a lot of effort to lobbying AI researchers to sign a declaration against such weapons. Both are cognizant of the threat of automation to jobs, though Tegmark mentions it only in passing in one section.

But for the most part, they are on totally different planes. As befitting a cosmologist, Tegmark is again thinking in grand terms: “Life 1.0” was life passively shaped by evolutionary forces, “Life 2.0” is us with culture shaping our brains, and “Life 3.0” will be us fusing with superintelligent machines. This is the stuff of movies like The Matrix or Space Odessey and here Tegmark displays his considerable talents at inventing fantasies: a whole chapter is devoted to various types of Matrix like future scenarios, some featuring benign AIs, others malignant. Life 3.0 starts with a parable of a future with a Hal like computer taking over the world. Perhaps a future career awaits Tegmark in the Sci-Fi movie industry! Tegmark also possesses great entrepreneurial talent – he has founded his own Future of Life Institute (FLI) devoted to these questions and secured a donation of $ 10 million from that other great entrepreneur Elon Musk who also likes to indulge in such speculations. There is an entire chapter in the book about the drama surrounding an event organized to announce the institute and the grant. One can see the need for hyperbole in such projects, but it borders on irresponsible to claim, as Tegmark has done, that AI is a more imminent existential threat than climate change.

Noted roboticist Rodney Brooks has warned of the “seven deadly sins of AI predictions” [5]. In particular he issues a warning about imagined future technology:

If it is far enough away from the technology we have and understand today, then we do not know its limitations. And if it becomes indistinguishable from magic, anything one says about it is no longer falsifiable.

Life 3.0 is guilty of several of these deadly sins.

Walsh’s concerns with AI risks are of a totally different sort. He is one of the sceptics about superintelligence and outlines his own argument why it may never ever be possible. He is not worried about superintelligent machines but rather super stupid machines with their bugs and failures and how we are reposing faith in them. He is worried about systematic biases in AI systems and consequences for fairness when they are entrusted with decision making responsibilities. And he is concerned about the consequences of automation on jobs and the economy. His own speculations at the end of the book are not all that implausible in the not too distant future: that you see your doctor daily (because it’s a computer) or that Germany loses a soccer match to a robot team!

The AI community has started taking the risks of AI seriously, for instance the recent work from Google [6,7]. These efforts are closer to the concrete down-to-earth approach of Walsh, for, as they write,

in our opinion, one need not invoke … extreme scenarios to productively discuss accidents, and in fact doing so can lead to unnecessarily speculative discussions that lack precision … We believe it is usually most productive to frame accident risk in terms of practical (though often quite general) issues with modern ML techniques.

This harks back to the wise words of Francois Jacob [8]:

The beginning of modern science can be dated from the time when such general questions as “How was the Universe created” … “What is the essence of Life” were replaced by more modest questions like “How does a stone fall?’ How does water flow in a tube?” … While asking very general questions lead to very limited answers, asking limited questions turned out to provide more and more general answers

References

[1] Max Tegmark, Life 3.0: Being Human in the Age of Artificial Intelligence, Knopf 2017.

[2] Toby Walsh, Android Dreams: The Past, the Present and the Future of Artificial Intelligence, C Hurst & Co Publishers Ltd, 2017.

[3] P. Domingos, The Master Algorithm, Basic Books 2015.

[4] K. Grace et al, “When will AI Exceed Human Performance: Evidence from the Experts”, arxiv 2017.

[5] R. Brooks, “The Seven Deadly Sins of AI Predictions”, MIT Technology Review, Oct 6, 2017.

[6] D. Amodei et al, “Concrete Problems in AI Safety”, arxiv. 2016.

[7] J. Leike et al, “AI Safety Gridworlds”, arxiv. 2017.

[8] F. Jacob, “Evolution and Tinkering”, Science, Vol. 196, Issue 4295, pp. 1161-1166, 1977.

 

 

 

 

 

[1] https://chess24.com/en/read/news/deepminSd-s-alphazero-crushes-chess

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s