It’s frustrating how often articles get posted at the NYT that engage in specious, uncritical thinking — especially when it’s in the service of the rich and wealthy. Today’s example is Six Charts That Show Just How Much We Need A.I. (gift link) by Steven Rattner. His argument is that we need AI, not just because it is innovation, but because it will drive innovation. I call shenanigans.
He’s primarily focused on growth. He links productivity growth with higher standards of living. Of course, that hasn’t been true since Reagan. In the 1980s, productivity growth has only gone to enrich the very wealthiest and wages have been stagnant. He can fudge his argument by selectively showing periods on his charts.
I really loved the graph cited above. Yes! All of those “Data entry keyers” should clearly get jobs as “Data Scientists”. Problem solved!
Where he really lost me (if he ever “had me”) was when he started talking about innovation. He describes how [AI] “saved me and my colleague, Will McGrew, many hours of prowling the internet to research this essay. That left us with time to do other tasks.” First, the only reason it could save you that much time is because the oligarchs wrecked Internet search to force people to use their stochastic parrot. It was way easier to find things on the Internet before they broke it. Second, if you outsource your thinking to a robot, then you undermine the most important path to innovation.
Innovation is frequently taking ideas from one domain and leveraging them in another. People can do this when they develop expertise in multiple domains. When you outsource your thinking to a robot, you lose most or all of the ability to do that. You can, perhaps, more deeply specialize in your own domain. But without multidisciplinary thinking, you’re less likely to have the transformational insights that drive true innovation.
Most laughably, was his admission that “Thus far, the gains from artificial intelligence have accrued not to workers, but to investors who have reaped soaring profits from the explosive increase in the value of companies in the A.I. industry.” But there are lots of ways to fix that, he assures us. Please. The only way that’s getting fixed is if people organize and fight back. Like they’re doing. Maybe you could help, dude, rather than whining that we need to let the corporations help us grow and innovate.
Where we may be in agreement is in the need for additional regulation. We need to shut down the A.I. companies and let the research be done in public laboratories until it’s demonstrated to be safe and effective to be used. We wouldn’t let pharmaceutical companies release their beta cancer treatments to the public. The AI companies should not be releasing these untested systems to see what havoc they can wreak on the public until we have safeguards in place, and the assurance that they’re safe and effective.
AI may, someday, be shown to be safe and effective. But we should fight these companies trying to jam it down our throats as if our lives depended on it. Because they just might.
