Artificial intelligence, or AI, is shaping up to be one of the most important topics of our time. There’s no question that you’re going to hear a TON more about it in the coming years. But like many complicated topics, misunderstandings float around in the media, leaving people confused. What follows is my attempt to clarify and distill down the most important aspects of AI that the public needs to know.
1. Extremely intelligent AI could be closer than you think.
When I was a kid back in the late 80’s and early 90’s, I remember thinking that “the future” was indeed a very far future—like hundreds of years away. Computers that talk and interact with us, or do our jobs for us, seemed like something I would see in the movies but never actually experience in my lifetime
But many technology experts across computer science, physics, philosophy, and other fields believe that surprisingly smart AI lies just around the corner. In one survey of computer science researchers, most of them thought there was a good chance of “high-level machine intelligence” by the year 2050, and a decent chance it will come by 2024. (They define “high-level machine intelligence” as an AI system “that can carry out most human professions at least as well as a typical human.”)
One window into this very near future is the impressive things AI can already do: Drive a car. Teach itself to beat 49 Atari video games. Beat any human at chess.
In fact much of the software you interact with everyday, from Google’s search engine to Apple’s Siri, uses a kind of limited AI (known in geek speak as “narrow AI.”)
2. AI will make many jobs obsolete.
From what I’ve read, this is the least debated prediction about AI. Even the most optimistic technology experts say that automation will inevitably eliminate certain jobs. This is because many simple and repetitive cognitive tasks can probably be done by software in the near future. Some professions currently on the chopping block: paralegal, medical technician, and many driving jobs—taxi, Uber, truck and delivery drivers.
Of course, new types of jobs could be invented, but there’s no telling whether we will have the political and economic will to make that happen.
3. Many smart people think AI poses a threat to humanity itself.
This is where it gets trippy—and a little scary.
If AI keeps getting smarter, eventually it will surpass our own intelligence, and then what happens?
In order to wrap your head around this, you must first disabuse yourself of the sci-fi movies you’ve seen. Forget about robot armies storming over the hills to fight humans in an epic battle of man vs. machine. Experts and common sense tell us that’s a highly unlikely scenario.
What concerns many experts, such as computer scientist Stuart Russell, is that AI could become so competent and powerful, that the slightest divergence between its goals and our own could be disastrous. The problem, Russell says, is that AI needs to be aligned with human values—at least as much as possible.
In other words, AI has to want what we want. Given that some experts think we might hit “high-level machine intelligence” by 2024, that doesn’t give us much time to sort this out.
Other luminaries who have expressed concern about the immediacy of this problem include Stephen Hawking, Bill Gates, and Elon Musk.
4. Many other smart people are optimistic about the future of AI.
Despite the risks, the possible benefits of increasingly intelligent AI are hard to deny.
In the long term, if progress moves in the right direction, AI could supercharge our mathematical and scientific research, helping us cure diseases or engineer robots to serve human needs.
Mark Zuckerberg has spoken publicly about his sanguine vision of AI’s near-term future, in which it could be used to diagnose diseases more accurately or drive cars more safely than humans do.
And former Wired editor Kevin Kelly also pushes back against AI fears, arguing that intelligence is not just “one thing.” Instead he sees AI as a disbursed phenomenon that could become a utility somewhat like electricity rather than a unitary superhuman entity.
5. AI is an incredibly complicated topic.
The physicist Richard Feynman once said of his field, “If you think you understand quantum mechanics, you don’t understand quantum mechanics.” What he meant was that quantum mechanics is such a strange, counterintuitive subject that it breeds confusion. I suspect this might also apply to the issue of AI.
Like many complicated topics, the media oversimplifies the AI discussion, making certain aspects of it misleading. Experts disagree on many questions within the field, a fact that often gets overlooked.
What’s sure though is that this is one of the most fascinating areas of science today.
To learn more, check out this mind-blowing article from Tim Urban called “The AI Revolution,” and this series of TED Talks.