Continuing my theme of doing things other than fret about Donald Trump, I have spent some time fretting about other existential threats to humanity. So, that’s healthy.
Specifically, I’ve spent the last half day thinking about the threat of alien invasions and runaway artificial intelligence. One of them you can consign to the bottom of your worry list; the other probably deserves a higher spot on the list, somewhere below Donald Trump but above death panels and “radical Islamic terrorism.”
The topic of alien invasions is the overt theme of the movie I saw last night: Life, directed by Daniel Espinosa and starring, among others, Jake Gyllenhaal, Rebecca Ferguson and Ryan Reynolds. Without giving away the plot, it explores the question of what happens when humanity encounters a lifeform that turns out to be smarter and more dangerous than it appears. Suffice it to say not all ends well for our gender- and ethnic-balanced crew aboard the International Space Station.
Despite the title of this post, the movie is not actually bad; it’s suspenseful and engaging. As I watched it, though, I was struck by how shitty the science was. As the investigators probed the alien lifeform, they repeatedly demonstrated all sorts of stupid, unrealistic practices. They let a single investigator engage in isolation with the lifeform to the extent he loses perspective. They do not do carefully measured experiments to determine both what sustains the organism and what kills it. When it demonstrates exponential growth and unexpected abilities, the researchers don’t react to this with caution but instead step on the accelerator. And, when things go wrong, they discover that their failsafe mechanisms are either non-existent or simply failures. Any epidemiologist or biologist working with potentially hazardous organisms would have been appalled.
The good news is that we’re not out scooping up biomass from other planets and bringing it back to Earth. There’s also every reason to think that the product of other evolutionary forces would not be particularly compatible with Earth’s. And, finally, there’s the fact that – despite the fact that we’ve been actively looking for decades, there’s very little sign of life – particularly intelligent life – outside of our little blue ball despite the fact that it’s a very, very big universe. This is known as the Fermi Paradox. My best guess is that you can put this issue way, way down on your list of things to worry about.
Which brings me to the other one, the existential threat of runaway artificial intelligence.
As I was driving home from the theater, it occurred to me that the movie was actually a commentary on the how we – not you or me, but some VERY smart people – are approaching the field of AI. As near as I can tell, we are using the same shitty scientific methods – the ones that would make any life science researcher cringe – to develop this technology. We have researchers all across the world laboring in secret, scientists who are less objective researchers and more would-be parents who are enraptured with the idea of strong AI or even the Singularity. Instead of running carefully controlled experiments and building in rigorous “kill steps,” AI is being deployed today in the real world – in Teslas, in fraud detection systems, in your washing machine, writing both press releases and news stories, in your favorite search engine, in the warehouses of your favorite retailer, as robo-calls and a thousand other ways. And, even though these creations are demonstrating unexpectedly rapid growth and ability (an AI-driven computer recent beat the world’s best Go players – widely considered an incredibly hard game – 60 games to none; a computer program performed a similar fear against some of the world’s best poker players), researchers are plowing onward at even faster rates.
This is perhaps not the smartest thing we’ve ever done. And, it’s not just me, your friendly blogger, who thinks so. Smart guys like Bill Gates and Elon Musk are worried about this. So are really smart guys like Stephen Hawking.
By way of fair disclosure, there are plenty of very smart people – Ray Kurzweil perhaps foremost among them – who believe the coming era of big AI will usher in an unprecedented era for humanity, giving us access to pretty much everything and an infinite lifespan to experience it. That seems like a better outcome, but this point of view is a little cultish and perhaps optimistic without hard, objective reasons. Life – whether artificial or otherwise – constantly finds ways to break out of whatever boxes it gets put into. Including the boxes we build.
If you’re inclined to read more on this, Vanity Fair coincidentally published a long interview with Musk on this topic. It is worth the 20 minutes or so it will take you and give you something to worry about instead of Trump.
There. Doesn’t that make you feel better instead of worrying about the latest cluster fuck from the White House? Next week, I’ll write about the threats of pandemics and global warming. Just call me Mr. Good News.