Fortunately for managers, building human capital takes a long time. Or at least it used to be: artificial intelligence is, among other things, a technology that accelerates learning and increases abilities. Millions of people now use large language models. Not all of them are flirting with their chatbots; instead, they discovered that with the help of AI, they could perform tasks they had never done before and quickly learn subjects that were previously thought to be inaccessible. What happens if you suddenly increase the rate of human capital accumulation? This is one of the challenges that AI poses for the business world as it tries to understand the value of this technology.
For a number of reasons, it seems strange to think of AI as a tool for augmenting human capital. Doesn't its usefulness lie in intelligent automation that makes hard-earned human knowledge unnecessary? Leading artificial intelligence companies talk about a future in which their systems replace workers en masse. Large companies that are currently integrating AI into their business are almost certainly thinking along similar lines. They have to do this because AI is expensive. Microsoft charges a per-user fee for its enterprise chatbot Copilot. If a large company with thousands of employees wants to purchase Copilot seats for its staff, it expects to invest many millions of dollars each year.
Will this “spending” lead to a corresponding return? The easiest way for a company to answer this question is to think about new products or staff cuts that could generate revenue or reduce costs, respectively. (They can be combined, of course.) In its new report on “enterprise” AI, published this week, OpenAI offers a series of case studies focusing on products that replace human labor. A typical example is an artificially intelligent voice agent useful for customer service calls; The company says one such agent is currently saving companies “hundreds of millions of dollars annually.”
All of this makes it seem like replacing workers is the logical end goal of enterprise AI. But it's important to note that, both conceptually and from an internal accounting perspective, large companies often have difficulty figuring out how to integrate new technologies. In the eighties and nineties, when IT departments were new, it was sometimes unclear how they could be justified internally. The IT department can spend millions every year on new computers, networking equipment, or productivity software. Did all this spending bring profit? How can you assess its value? If a large corporation installed a mainframe computer, it could replace some of the accountants. If an IT manager wanted to explain to her boss why computers are so important, the simplest thing she could say is that they can replace the typing pool.
However, over time it became clear that the costs and benefits of information technology far exceeded what could be explained in this way. Modern companies have reorganized themselves around computers; In this new world, the goal of IT departments was not to replace computer-dependent workers, but to improve their efficiency. Workers began to demand more from their IT departments. In a development known as “consumerism,” the tools used by tech-savvy employees at home, such as smartphones, have become more advanced than those provided at work; employees who wanted to do more began demanding updates. As a result, today, when IT “spending” is proposed, no one insists that the investment be anything as crude as replacing workers. An important question is whether new investments help existing employees perform their tasks and keep pace with their competitors in other firms.
The idea that the best use of AI (perhaps the only beneficial use) is direct replacement of workers combines two schools of thought: one stemming from speculation about the future of AI, and the other from the short-term balance sheet thinking that is likely inevitable when companies explore new technologies. Meanwhile, this is deeply at odds with the experience many of us have when actually using AI. Huge numbers of people pay for accounts with OpenAI, Anthropic and other companies because they believe AI makes them more capable and productive. From their point of view, this is a human capital multiplier. If you have a clear idea of what you want to achieve –write softwareanalyze research, diagnose a diseaserenovate something in your home – artificial intelligence will help you do it faster and better. Companies today spend a lot of money on training their employees; even highly qualified employees are being sent to online seminars and expensive retreats in the hope that they will return improved. Let's say AI makes some employees five or ten percent more knowledgeable and capable. How much should a company pay for this cognitive boost?
