This month, the National Bureau of Economic Research (NBER) released a study that confirmed what is increasingly becoming obvious (except to the stock market). An article about the study starts with this sentence: “A National Bureau of Economic Research study surveying nearly 6,000 executives across the US, UK, Germany and Australia found that more than 80% of firms reported no impact from AI on employment or productivity over the past three years, despite widespread adoption and billions in investment.”
I want to unpack that statement:
1.     Companies in major companies on three continents are “widely” adopting AI, and spending billions of dollars on the technology.
2.     Despite that fact, they have no plans to reduce employment. Of course, this contradicts the prevailing wisdom that AI will result in a lot of white collar workers losing their jobs. There are two groups of people saying this will happen: a) People that are genuinely worried about the effects of large-scale permanent unemployment on society and b) Executives of companies with a substantial stake in AI’s success, who are making that point to advertise their AI products as a great vehicle to cut costs by laying off staff. It seems both groups are wrong.
3.     AI isn’t impacting productivity, despite the billions being spent on it. This is why companies aren’t laying off workers, since absent productivity gains, doing that could cause companies to miss product delivery deadlines, cut services to customers, fail to meet legal or regulatory obligations, etc.
Here are some more data points I’ve run across:
A.    This week, the AI Supremacy blog put up this post that made some interesting points:
·       Regarding a PwC survey report: “Despite $30–40 billion in enterprise investment into GenAI, this report uncovers a surprising result in that 95% of organizations are getting zero return.”
·       Regarding the same report: “Despite $30–40 billion in enterprise investment into GenAI, this report uncovers a surprising result in that 95% of organizations are getting zero return.”
·       Regarding a January Gallup survey: “In industries such as technology where AI use has been most prevalent, growth in total users shows signs of leveling, with gains found primarily among those already using AI.”
·       “Outside of the AI adoption stories and lagging data for people and workers, Generative AI is not seeing evidence of new categories of jobs, boosts in productivity or ROI from Enterprise, companies or SMBs at any scale.”
·       “In 2024 and 2025, most AI pilots at companies didn’t lead anywhere.”
B.    In November, The Economist, in an article about a survey recently conducted by the Census Bureau, reported “..the employment-weighted share of Americans using AI at work has fallen by a percentage point, and now sits at 11% (see chart 1). Adoption has fallen sharply at the largest businesses, those employing over 250 people. Three years into the generative-AI wave, demand for the technology looks surprisingly flimsy.”
C.     Last fall, a report from Stanford University touted the huge investment boom in AI, but only addressed the question of whether businesses were using AI in one paragraph: “Most companies that report financial impacts from using AI within a business function estimate the benefits as being at low levels… most (companies) report cost savings of less than 10%...the most common level of revenue increases is less than 5%.”
Of course, this doesn’t mean generative AI is a bust. Instead, it means (as I discussed in this post) that AI is like every other revolutionary technology: factories, the steam engine, railroads, the automobile, electric lighting, semiconductors, microprocessors, desktop computers, the internet, the world wide web, social media, etc. Adoption requires substantial changes in how businesses operate, physical infrastructure, training availability, etc.
None of these changes occur in just three years (which is how long ChatGPT has been available). Five years is the minimum for substantial uptake to happen, and ten years is probably much more realistic. Unfortunately, it seems a lot of company valuations seem to assume a much shorter time frame.
One big difference between the AI technological revolution and previous revolutions is that the necessary infrastructure is being rolled out as quickly as possible, far ahead of the business demand for it. Of course, this is because there is already huge demand from hundreds of millions of individual consumers worldwide – the only problem is those consumers aren’t paying for what they’re using and probably will never be persuaded to pay for it.
However, the data centers that are popping up all over aren’t being built for free, even though what they produce is being given away for free. Why are the AI companies investing so much in something that isn’t bringing in much revenue at all (I heard three years ago that Microsoft was opening a new data center somewhere in the world every three days. Given how much more Microsoft is spending on AI now than three years ago, that number is probably down to two days or even one day)?
Obviously, these companies think the tsunami of business spending on AI will hit any day now; they want to make sure they’re ready for it. But they’ll be standing on the beach wearing their inflatable ducky for quite some time. I hope they brought a sandwich with them. Â
Tom Alrich’s Blog, too is a reader-supported publication. You can view new posts for two months after they come out by becoming a free subscriber. You can also access all of my 1300 existing posts dating back to 2013, as well as support my work, by becoming a paid subscriber for $30 for one year (and if you feel so inclined, you can donate more than that!). Please subscribe.Â
If you would like to comment on what you have read here, I would love to hear from you. Please comment below or email me at [email protected].