Is AI actually taking jobs? Spoiler alert: the data suggest it's not happening yet, despite all the anxiety out there.
On the SuperDataScience Podcast, I’m often covering the latest, most exciting AI advances. There’s often a dark flipside to that in many of our minds, however! Maybe you've tried ChatGPT or Claude and thought, "Wow, this thing can do a lot of what I do at work." You're not alone. Earlier this year, global Google searches for "AI unemployment" hit an all-time high. In tech hubs like London and San Francisco, "How long do you reckon you have left in your job?" has become standard water cooler conversation.
With AI models now capable of writing detailed reports, creating videos on demand, and tackling increasingly complex tasks with fewer hallucinations, it's natural to wonder if we're all about to become surplus to requirements. But here's the thing – when we actually dig into the data, the AI jobs apocalypse that everyone's worried about? It's nowhere to be found.
Let me walk you through some of these data. First, there's a recent paper by Carl Benedikt Frey and Pedro Llanos-Paredes from the University of Oxford that's been making the rounds. They suggest there's a link between automation and declining demand for translators. Sounds scary for translators, right? But here's where it gets interesting – official American data show that employment in interpretation, translation, and related fields is actually up 7% year-over-year. That's not a decline; that's growth!
Then there's the Klarna example that's been cited everywhere. The fintech company made headlines boasting about using AI to automate customer service. But guess what? They're now doing a complete about-face. Sebastian Siemiatkowski, Klarna's CEO, recently stated, "There will always be a human if you want." So much for the complete automation of customer service.
Separately, some folks have been trying to find evidence of AI job displacement in macroeconomic data. One metric that's gotten attention is the unemployment ratio between recent college graduates and the overall American average. Young graduates are currently more likely than the average worker to be jobless, and the theory goes like this: these young graduates typically do entry-level knowledge work – e.g., paralegal tasks, making PowerPoint slides at management consultancies – exactly the kind of work that generative AI excels at. So maybe AI has eliminated these jobs?
But when you actually examine the data carefully, this narrative falls apart as well. The relative unemployment rate for young graduates started rising back in 2009 – that's fifteen years before ChatGPT burst onto the scene. And their actual unemployment rate? It's around 6%, which is still quite low by historical standards.
And what of data that directly address whether AI is hitting white-collar workers – the folks everyone assumes are most vulnerable. We're talking about people in back-office support, financial operations, sales, and similar roles. If AI were decimating these jobs, we'd see it in the employment statistics, right? Well, we don’t. Over the past year, the share of employment in white-collar work has actually risen slightly. Let me repeat that – it's gone up, not down. American unemployment overall remains low at 4.2%, and wage growth is still reasonably strong. It's really hard to square strong wage growth with the idea that AI is causing labor demand to plummet.
And this isn't just an American phenomenon. Earnings growth across much of the rich world – Britain, the euro area, Japan – remains robust. In 2024, the employment rate for OECD countries hit an all-time high. That's the share of working-age people who actually have jobs reaching unprecedented levels at the exact moment when AI capabilities are exploding.
So what's going on here? There are two competing explanations, and both are probably true to some extent.
First, despite all the breathless announcements about companies revolutionizing their operations with AI, actual adoption for serious production work remains surprisingly low. An official measure suggests that less than 10% of American companies are using AI to produce goods and services. Think about that – we're in the midst of what many call an AI revolution, and 90% of companies aren't even using it for real work yet. (A big opportunity for all you listeners out there…)
The second explanation is that when companies do adopt AI, they're not firing people. Instead, AI is helping workers do their jobs faster and better. It's augmenting human capabilities rather than replacing humans entirely. This makes a lot of sense when you think about it. Most jobs involve a complex mix of tasks, many of which require human judgment, creativity, emotional intelligence, or physical presence that AI can't replicate.
Now, I want to be clear – I'm not saying AI will never impact employment. Technology has always changed the nature of work, and AI will be no different. But the data suggest we're not facing an imminent jobs crisis. Instead, we're in a phase where AI is creating opportunities for those who learn to work with it effectively.
Here's my take: rather than panicking about AI taking your job, focus on learning how to use these tools to enhance your capabilities. The most valuable workers in the coming years won't be those who compete against AI, but those who collaborate with it. Experiment daily with the latest LLMs (my favorite for general work today is Anthropic’s Claude Opus 4 though I reach for OpenAI’s Deep Research for my most complex and important requests; listen to Episode #870 for more detail) so that you understand what AI can (and can't!) do well and figure out how to integrate it into your workflow.
The bottom line is this: despite all the anxiety, the data show that AI isn't causing mass unemployment. Employment is high, wages are growing, and companies are still hiring. Yes, the nature of work will evolve, but that's been true throughout history. The key is to evolve with it.
So take a deep breath and smell the opportunity: start thinking about how you can use AI to become even better at what you do. These are exciting times, and with the right mindset, you can thrive in the age of AI rather than just survive it.
The SuperDataScience podcast is available on all major podcasting platforms, YouTube, and at SuperDataScience.com.