Imagine a world where artificial intelligence promises to revolutionize everything, but instead of lifting us all, it might just burst like an overinflated balloon—leaving economies in tatters. That's the chilling reality economist Jason Furman warns us about in his latest insights on potential AI bubbles. But here's where it gets controversial: Is the real danger in the tech itself, or just how much we're betting on it financially? Dive in as we unpack the key takeaways from his Bloomberg Q&A, and see if you agree that AI's future hangs by a thread.
Jason Furman stands out as one of the more relatable figures in the world of economics. He's a respected professor at Harvard University, where he dives deep into policy and economic trends (you can learn more about his work at https://www.hks.harvard.edu/faculty/jason-furman). Before that, he served as the chairman of the White House Council of Economic Advisers during President Barack Obama's administration—a role that gave him a front-row seat to major economic decisions (check out the announcement here: https://obamawhitehouse.archives.gov/blog/2013/06/10/president-obama-nominates-jason-furman-chairman-council-economic-advisers). Back in October, Furman appeared on the podcast hosted by conservative New York Times opinion columnist Ross Douthat (available at https://www.nytimes.com/2025/10/23/opinion/ai-bubble-economy-bust.html), and in that chat, he voiced just a single concern about AI. So, what led to this sudden surge in worries? Let's explore.
And this is the part most people miss: Furman kicks off by emphasizing he's more concerned about a financial valuation bubble than a technological one. At first glance, this seems like a subtle nuance—suggesting the underlying tech could be groundbreaking, but the companies investing in it might be priced way too high, making the overvaluation the true threat. Yet, his follow-up comments blur the lines, hinting that both aspects deserve equal scrutiny. He explains that for these sky-high valuations to make sense, two crucial elements are needed: the technology must perform exceptionally well, and it has to generate real profits. The risks? We could hit a wall with diminishing returns, where the scaling laws—those patterns that have driven AI progress so far—stop applying as effectively. Moreover, not every scaling improvement translates into economic gains. For instance, picture a computer microchip doubling in speed; does that mean you churn out twice as many emails or Word documents? Often not—much of that power ends up as untapped potential, a kind of excess capacity piling up. And this could mirror what's happening in AI, even if it continues to follow those scaling patterns.
This description eerily mirrors one of the year's biggest AI headlines (read more at https://gizmodo.com/it-took-just-24-hours-of-complaints-for-openai-to-start-bringing-back-its-old-model-2000640912). When OpenAI unveiled its GPT-5 model in August, it might have been a technical upgrade, but users of ChatGPT didn't perceive a big enough benefit to outweigh their frustrations. Essentially, OpenAI swapped out the model people relied on like a friendly companion, and it didn't magically become warmer, wiser, or more engaging. Instead, it felt like just more excess capacity—powerful, perhaps, but not immediately valuable.
Still puzzled about the difference between a tech bubble and a valuation bubble? You're in good company; Bloomberg's interviewer, Shirin Ghaffary, admitted the same. Furman provides more clarity, pointing out that beyond valuations, we're pouring hundreds of billions into real-world infrastructure each year—think data centers, energy sources, and related expenses—and this represents genuine, tangible activity. He draws a parallel to the dot-com boom, when massive investments went into internet infrastructure. But he adds a caveat: The real red flag would be if all this spending doesn't ultimately boost productivity. Currently, AI is mostly influencing the demand side of our economy, not yet the supply side.
He elaborates further: Our US economy isn't operating at full steam; it's more like running on just one cylinder right now. These observations are crucial for understanding how mainstream economists view AI today. When Furman mentions AI being on the demand side, it might sound odd—after all, how often do you personally crave AI in your daily life? For many of us, the answer is rarely, if ever. But he's not referring to personal desires; instead, imagine the global economy as a vast, often underutilized hardware store like Home Depot. AI on the demand side acts like an enormous, insatiable shopper, snapping up drills, bags of cement, and ladders in bulk to keep the store thriving temporarily. However, AI can't remain the sole big buyer forever. The structures it builds with these resources must generate enough economic momentum to attract even more customers—ideally, more than ever before—so everyone can create and innovate too.
Reflecting on this year's ChatGPT episode, while consumer interactions are a key part of AI's appeal, Furman argues they're not the primary driver for long-term economic growth. He also dismisses fears that AI will drastically reduce jobs through efficiency gains, calling it an unlikely major risk. "At every point in time that people have thought that in the past about this employment question, they've been wrong," he notes, referencing historical mispredictions. Instead, Furman's vision of AI's future is hazy: People and businesses will adopt it gradually, discovering new applications year by year, testing them thoroughly before full implementation. Different sectors and companies will catch on at their own paces, rather than a sudden, explosive transformation. Of course, this is just his educated guess, with the huge disclaimer that anything could unfold unexpectedly.
You might find this reassuring or not, but in my take, Furman is essentially predicting that AI will prove genuinely beneficial—though we don't yet know when or for whom. This isn't a wild stretch; think of how past technologies like the internet evolved unpredictably. The truly alarming aspect? For the economy to thrive, this prediction simply must hold true. But here's where it gets controversial: Some argue that AI's slow rollout could be intentional corporate strategy to avoid bubbles, while others fear it's a sign of overhyped potential that might never materialize. Do you think Furman's optimism is warranted, or are we ignoring red flags? Share your thoughts in the comments—do you see AI as a bubble waiting to burst, or a transformative force? Let's debate!
