The anti-human temptation of AI

If you’ve been watching the Olympics, chances are you’ve noticed the steady parade of strikingly tone-deaf AI advertisements filling the commercial breaks. One, for Google’s Gemini, features a young girl striving to imitate her idol, track star Sydney McLaughlin-Levrone. In a voice-over, her father says that she “wants to show Sydney some love,” and asks Gemini to write a fan letter to the Olympic hurdling champion. The ad soon went viral for all the wrong reasons.

“I flatly reject the future that Google is advertising,” wrote one media scholar, Shelly Palmer. “I want to live in a culturally diverse world where billions of individuals use AI to amplify their human skills, not in a world where we are used by AI pretending to be human.”

Other AI ads, however, were only slightly less overt in their anti-humanism. In one, for Microsoft Copilot, an employee realizes he hasn’t adequately prepared for an important work meeting. He quickly pulls up Copilot and asks the software to summarize 150 pages worth of material into key points for his presentation. His face lights up as he imagines wowing his co-workers with the snazzy charts and slides.

It is not hard to believe that many of us, in such a pinch, might ask technology to help bail us out. What does beggar belief, however, is that any of us should feel proud of it. The man in the ad, like the workers in dozens of similar ads for AI tools this year, seems to radiate with a sense of joy and accomplishment in his work. But we all know that those emotions depend on a sense of real achievement. Indeed, in many pursuits, the pride we feel in an achievement is directly proportional to the amount of blood, tears, toil, and sweat it took to get there. Riding a gondola may give you the same breathtaking views as scaling a mountaintop on foot, but chances are the latter will provide far richer and more lasting memories. In some pursuits, we recognize that taking shortcuts isn’t just lame; it should be a source of shame.

Indeed, we used to have a word for having someone else write a presentation for you and passing it off as your own: cheating.

It is particularly ironic that such ads should fill the airwaves in between showcases of real human beings pulling off truly astonishing feats of real human achievement at the Olympic Games. When we discover that some record-breaking swim or jump was fueled by artificial performance enhancers, we are rightly scandalized. How much more so if we discovered the swimmer wasn’t human at all! “But why?” the AI enthusiasts might ask. “If it’s amazing to watch someone long jump 30 feet, wouldn’t it be even cooler to have a world where robotically-assisted athletes jumped 50 feet?” Google’s marketing team might think so, but the rest of us know better.

The problem is that although we may know better, when it comes to action we tend to follow the path of least resistance. AI can be a very powerful tool, doing hours worth of mind-numbing grunt work in seconds so we can focus on higher-level thinking, creating, and communicating. But it can be tempting to let it start doing that for us as well. We begin by letting the AI correct our spelling, and soon find ourselves letting it draft entire emails. At first, we might ask it to summarize some articles for us so we can save time on research and focus on thinking and writing, but before long—especially if we find ourselves in a pinch—we may ask it to do the thinking and writing too, so we can focus on … enjoying more leisure time, perhaps? What begins as a supplement can quickly become a substitute, if we’re not careful. Indeed, as Joshua Mitchell has argued, this is the temptation when it comes to so many of our technologies today, from pain medicines to online shopping.

Technology has always been a two-edged sword. On the one hand, if well-conceived and well-used, it can help unlock hitherto unimagined human potential, allowing our minds and bodies to achieve extraordinary feats of creativity, power, and discovery, and steadily multiplying our capabilities. On the other hand, it can easily become anti-human, offering us pleasure or results without the trouble of thought or effort. Not only does the lack of effort take something away from the pleasure or achievement, but it deprives us of the opportunity for growth, causing our capabilities to atrophy. If we are too careless in letting Copilot take the controls, we may find ourselves hurtling toward a future not unlike that of Pixar’s dystopian tale WALL-E, in which the surviving humans, glued to their screens and their Slurpee straws, became less authentically human than their robots.