Corporate AI Isn't Actually Making (or Saving) Very Much Money

5 days ago 7 Back

Hands on a laptop

Credit: Manuel Breva Colmeiro/Getty Images

Pretty quickly after the introduction of chatbots like ChatGPT, it became clear that the MBA-driven quest to replace workers with AI agents simply was not going to work, at least in the near future. Alright, they said, but AI will still increase productivity and have a huge impact on the corporate and national bottom line.

Well, now a high-profile survey is calling even that idea into question. The National Bureau of Economic Research surveyed 6,000 CEOs, CFOs, and other C-Suite decision-makers from large corporations. They were from all over the English-speaking world, both start-ups and legacy corporations.

The survey found that the vast majority of respondents had yet to see AI have any meaningful impact on their business, beyond the direct impact of its own subscription cost. Almost 90% of the leadership surveyed felt they'd seen no impact on employment or productivity over the past three years. (Interestingly, executives themselves don't report much first-hand use of the tech; while a majority reported using AI regularly, that regular use chalked up to only 1.5 hours per week.)

Many of the business leaders did remain bullish on the future, however. In particular, they believed that over the next three years, AI would drive a 0.7% decrease in employment. Individual employees were also surveyed on this question, and said they predicted a 0.5% increase in employment.

There was more agreement on productivity. Over those same three years, most respondents predicted "sizable impacts... forecasting AI will boost productivity by 1.4%, [and] increase output by 0.8%."

So, we have to ask: Why are they so sure that the next three years will be superior to the last three?

Corporate meeting

Credit: Nitat Termmee/Getty Images

After all, companies' AI costs have already risen steadily, with corporate spending on AI increasing by more than 100% year over year. These services show every sign that they will continue to rise in price, most especially because current prices still don't allow providers turn a profit.

The numbers are fuzzy, of course, but when even OpenAI seems to be hemorrhaging money, something's got to give. Third parties like the uber-valuable CoreWeave were previously seen as a rare safe spot within the emerging ecosystem of companies, but recently, even those middlemen have been revealed to have highly questionable fundamentals.

Yet, as Futurism reports, the cultural belief in AI's future impact is ironclad; when asked what concerns them most about the current lack of returns on AI investment, many said they thought they needed to spend more on AI. Much of the modern business world, both in and out of pure tech, is currently so indebted, over-leveraged, and in danger of overnight disaster that the overall mentality seems to be, "screw it."

It's a simple decision matrix: Either a company invests in AI or it doesn't, and either AI provides totally necessary returns for users or it doesn't.

If a firm does invest and AI does pay off, then it's in an elite league of companies that will survive and thrive in the next generation. If it invests and it doesn't pay off, then it's taken a crippling amount of debt for no increased income—but so have its competitors. If a firm doesn't invest and AI ends up paying off, then the firm is deader than a coffin nail; if it doesn't invest and AI doesn't pan out, it's got a bigger war chest left over than competitors who did invest.

Out of these scenarios, half say, "Definitely do invest," while the other half say, "Maybe don't." That's two "definitely" results versus two maybes, as evaluated by a group that often expects to personally move on from company losses as though they never happened.

The cavalier attitude about whether this is all even supposed to be justifiable also has to apply to the AI industry itself, of course. This leads to wonderfully pickled thoughts like those jarred up in this headline from the Harvard Business Review: "AI Companies Don't Have a Profitable Business Model. Does That Matter?"

That's really the most galling thing about the slow-motion nature of the oncoming AI bubble-burst: When it does finally happen, the implosion of most of the Western economy will be set to a chorus of millions chanting, "I told you so."