In the argument between Mark Zuckerberg and Elon Musk, it’s hard to know which side to join. Both of them are right. Or, if you like, both of them are wrong.
Musk is wrong to worry about artificial intelligence (AI) being a threat to humanity, so I agree with Zuckerberg. And Zuckerberg is wrong to dismiss all concerns about AI, so I agree with Musk. But neither of them are worrying about the right things.
AI is transforming almost every aspect of our lives, from the workspace to the political arena. You can’t open a newspaper today without reading a story about some impressive advance in AI.
Are machines taking over people’s jobs? Are algorithms having an impact on political debate? Will robots transform warfare? Are we sleepwalking into some dystopian future?
We’re working on ‘AI safety’
First, let’s put to rest Elon Musk’s worry. The machines aren’t about to take over the world anytime soon. Those of us working on building intelligent machines appreciate how much of a challenge remains. We’re not going to wake up anytime soon and discover the machines are in charge.
Most of my colleagues working in AI estimate it is at least 50 years before we can build machines as smart as humans. And when we do, it’s not inevitable they’ll be able to make themselves even smarter still.
So, there is plenty of time to ensure the machines are working in our best interests. And there’s a healthy community of researchers working on the topic of “AI safety” to ensure that outcome.
But that doesn’t mean we can simply put our feet up and wait for the bright future. There’s a lot to worry about. Some AI is smart, some is stupid. We’re starting to give responsibility to algorithms that aren’t actually very intelligent.
Joshua Brown discovered this to his cost in May last year. He was immortalised as the first person killed by their autonomous car. His Tesla was driving down the highway in “autopilot mode” when it hit a truck turning across the road. Mr Brown had too much faith in the technology.
Another worry is the impact AI is having on political discourse. When millions of Donald Trump’s Twitter followers are robots, you have to worry if human voices are being drowned out by computers. If the news you see on Facebook is decided by algorithms, who decides on the biases in these algorithms?
A third worry is the impact AI will have on the workforce. There’s no fundamental law of economics that requires new technologies to create more jobs than they destroy, which has been the case so far. There are more people working today than ever, and unemployment is at historically low levels.
There were 50 years of pain after the Industrial Revolution
But this time could be different. In the Industrial Revolution, machines took over much manual labour but left us with many cognitive tasks. In the AI revolution, machines will take over many of these cognitive tasks. What is left for us?
The Industrial Revolution offers us a good historical precedent for dealing with change like this. Before the industrial revolution, many people worked out in the fields. After the Industrial Revolution, machines took over many of these jobs. And new jobs were created in offices and factories.