If you're struggling to sort the AI hype from reality, you're not alone. The seemingly breakneck pace of AI development makes it tough to sort headlines from fantasy, with a constant flood of new products and incremental improvements to old ones combining into a rhetorical mess.
Arguably the main economic risk of developing artificial intelligence — or its biggest draw, if you're a business owner trying to pad your bottom line — is the prospect of automating jobs.
Whether or not AI is currently taking a meaningful number of people's jobs, though, has been exceedingly difficult to nail down. On the one hand, the US job market has taken a nosedive in recent months and the "laptop workers" most vulnerable to AI seem to be getting hit especially hard.
On the other hand, even the most advanced AI still struggles to perform alongside humans, let alone replace them. A recent MIT study found that AI initiatives are failing to deliver expected revenue returns at 95 percent the companies that roll them out.
But while there's plenty of reason to doubt the AI industry's extravagant claims, there's also reason to worry that the tech is already eating into the labor market.
For instance, a recent survey of AI and labor data by a team of researchers at Stanford's Digital Economy Lab uncovered some of the first comprehensive evidence that the AI industry really is throwing the job market into flux.
The scholars compiled three years of payroll records on millions of US workers at tens of thousands of businesses, allowing them to identify long-term trends according to jobs and age groups.
Their first finding was a dramatic decline in employment for entry-level knowledge workers aged 22 to 25 years old, whose occupations are at the highest theoretical risk for automation — a metric called "AI exposure." These are workers in office gigs whose day-to-day tasks have a lot of crossover with AI functions, like software engineers, service workers, and marketing professionals.
By comparison, older workers in those fields saw their headcount either stagnate or increase slightly.
But what gives the study some extra bite is its observation of 22 to 25 year olds in non-AI-exposed job fields — like industrial workers, nurses, or retail supervisors — who saw their headcounts grow over the past few years.
... continue reading