Let’s circle back one more time to Treasury Secretary Steve Mnuchin’s extraordinary comment that artificial intelligence is “not even on my radar screen” and won’t affect the job market for 50 to 100 years. What was he smoking? And please keep it away from us, okay?
The double whammy of AI and robotics, what observers are calling “the fourth industrial revolution,” is certainly on the rest of the world’s radar, and already having an impact on transportation, manufacturing, retail, medicine, education — everything. We can’t know how this wave of change will play out; scenarios range from utopia to doomsday, and we’re already beginning to live them.
Here’s a roundup of some of the issues and data points you might want to review. If you know Mnuchin, pass them on to him.
BlackRock Lets Software Pick Its Stocks
Right in Steve Mnuchin’s backyard (he’s a former Goldman Sachs exec and hedge fund manager), the AI algorithms are coming for financiers’ jobs. At BlackRock, the largest private equity fund in the world, software-driven stock-picking is rapidly displacing human-directed investment decisions, and putting fund managers out to pasture (The New York Times).
“Actively managed” funds helmed by actual people have higher costs and profit margins than passive index funds and other rules-based investments. But investors are voting with their feet and choosing the lower-cost, machine-based alternative. That might be because the actively managed funds usually perform more poorly (at BlackRock, only 11 percent of the active funds beat their indexes).
Robots Really Do Want Your Job, Economists Find
Does automation actually replace human laborers, or just make them more productive? A new study of industrial automation by two economists at MIT and Boston University finds that, from 1990 to 2007, for every robot deployed in a factory, three to six human jobs disappeared (Quartz).
Yikes! The impact can be seen in the most concentrated way in the auto industry. During previous industrial transitions, similar job losses turned out to be short-term dislocation costs — painful but not permanent. But who can guarantee that this time around will go the same way?
Is There an AI In the House?
Neural networks are getting better and better at making medical diagnoses. Will they gradually replace human diagnosticians like radiologists and dermatologists? Siddhartha Mukherjee pursues this question in a deep, thoughtful New Yorker piece.
A study led by Sebastian Thrun found that it didn’t take long for a well-designed “deep learning” system to become better than human doctors at identifying melanomas. Thrun doesn’t fear for the medical profession; he thinks the neural net diagnosticians will “augment” human doctors rather than replace them. But computer scientist Geoffrey Hinton is less sanguine: “It’s just completely obvious that in five years deep learning is going to do better than radiologists. It might be ten years…. They should stop training radiologists now.”
AI doctors will get better and better at figuring out how to identify patterns faster than people, Mukherjee concludes. But surpassing us at “how” doesn’t mean they will be able to replace doctors, because the most important questions the human MD asks is “Why?”
Why Elon Musk Is Very Afraid
Tesla founder Elon Musk is worried that AI might not only replace individual workers but run amok and destroy the entire human species. In Vanity Fair, Maureen Dowd chronicles the evolution of the billionaire entrepreneur’s thinking on how AI could go bad.
In a 2014 MIT speech in, Musk declared, “With artificial intelligence, we are summoning the demon.” But surely, if we’re worried that some future AI is going to evolve out of our control, we can just build in an “off” switch? Won’t help, according to Musk: “I’m not sure I’d want to be the one holding the kill switch for some superpowered A.I., because you’d be the first thing it kills.”
Today Musk is taking more of an “If you can’t beat ’em, join ‘em” approach — promoting the idea of cyborgish “neural lace” devices to create a direct brain-computer interface that would give future AIs root access to our brainstems. He is also pushing for an open model of AI: If we distribute the technology widely, we can make sure it doesn’t end up “concentrated in the hands of tech or government elites.”
Letting an imperial few monopolize such power is simply too risky: “It’s great when the emperor is Marcus Aurelius,” Musk says. “It’s not so great when the emperor is Caligula.” Or, perhaps, Trump.
Trump’s AI Blind Spot
So we’re at a crossroads for our economy and our lives. Making smart choices about whether and how to regulate or restrict this powerful new technology could determine the fate of our children and even or species. Too bad we in the U.S. took this moment to elect a president who “seems totally oblivious to these trends,” as Timothy B. Lee puts it (Vox).
It’s not just Steve Mnuchin. Self-driving cars are already prowling many of our cities’ streets, yet when President Trump talks about the auto industry, he doesn’t even mention the phenomenon. Trump is single-mindedly focused on foreign competitors stealing U.S. jobs, even as automation is almost certainly a larger and longer-lasting trend.
“Over the next four years, federal policymakers are going to play a crucial role in determining how AI-related breakthroughs affect the American economy and society, whether Trump realizes it or not,” Lee writes. Is there time to give the president one of Musk’s neural-lace implants?