For AI programmer Anthony Levandowski, the idea that a future superintelligent machine would be god-like is neither a metaphor nor a science fiction trope. It’s a goal, apparently.
Levandowski founded a self-driving truck company that was bought by Uber in 2016 (and is now the focus of an intellectual property lawsuit between Uber and Google robocar successor Waymo), so he knows a thing or two about artificial intelligence.
His non-profit religious organisation, the Way of the Future, reportedly seeks to “develop and promote the realization of a Godhead based on artificial intelligence and through understanding and worship of the Godhead contribute to the betterment of society”. Although the Way of the Future was registered in September 2015, its existence was only recently revealed in the .
Advertisement
The idea of working on an AI-based “Godhead” may be less outlandish than it seems. Behind it could lie a modified version of Pascal’s wager: the notion that, based on possible gains and losses, it is rational to believe in a god even if that deity might not exist. If superintelligent machines are possible, then they would be likely to appreciate you working to bring them about and reward you for showing faith. If they aren’t possible, then you have still made the world a better place by advancing computer science.
Levandowski, who has yet to publicly spell out the details of his movement, is by no means alone in believing that the emergence of superpowerful AI could change the world. Technologists such as Ray Kurzweil at Google see the advent of self-improving artificial intelligence as a globally transformative “singularity“: an explosive advance in technology that leaves us humans playing second fiddle. Kurzweil says it would forever reshape every aspect of who we are and how we live, and in ways we can’t begin to imagine.
Act of faith
While lots of people will dismiss Levandowski’s reported goal as odd, if you subscribe to the idea that a true AI is possible, it isn’t that outrageous. In many ways, it is a struggle to describe the potential capabilities of a superintelligent entity in ways that don’t sound religious.
What The Way of the Future does is take the tacit religiosity of the singularity concept and make it explicit. But will all be as benign as it seems to suggest?
Levandowski isn’t the only person working towards a singularity-strength artificial intelligence. If this is achievable, we are therefore likely to see a multitude of them – a pantheon, if you will – rather than a single digital deity. As these super-AIs will have been programmed by groups with varied goals and ethics, it isn’t difficult to imagine rivalries between the resulting machines. Fights between gods rarely work out well for mere mortals, at least according to classical mythologies.
There is another reason why we may all need to take talk of an AI godhead more seriously: the philosophical concept called . This asserts that anyone who knows that a superintelligent, godlike AI is possible but doesn’t work to bring it about will be considered by said AI as having effectively opposed its creation, and therefore be deemed an enemy.
According to this concept, not knowing of either the possibility or the implicit responsibility to bring about super-AI may give you a get-out clause come judgement day. Unfortunately, now that you know about Roko’s Basilisk, you must either help to “” or be doomed should it come about. Sorry about that. Maybe Levandowski got that memo too.
Read more: The God issue: New science of religion
Article amended on 4 October 2017
It was Uber that bought Levandowski’s self-driving truck company
