Better living through AI?

We’ve all had the experience—probably more times than we care to admit—of yelling into a phone receiver, “Just give me a human being!” after what feels like the umpteenth set of automated options on a customer service line. However, many of us have probably also found ourselves almost missing the clear and soothing tones of the automated assistant after being transferred to a customer service agent with an almost indecipherable foreign accent and, if possible, even less empathy and imagination than the robot we were previously speaking to. “Transfer me to your manager!” we may bark out if we’re having a particularly bad day.

New artificial intelligence companies like Sierra are promising to spare you that bad day, using generative AI to replace bad robots—and robotic humans—with genuinely helpful robots. The company, already valued at $4 billion, is one of several AI firms promising to revolutionize the global customer-service industry, trimming much of its $120 billion in annual labor costs and making life better for consumers at the same time.

It’s not hard to see the appeal. New generative AI bots, unlike older automated assistants, are capable of processing ordinary human speech, not just one-word menu item responses. And they are capable of reading and assimilating vast troves of technical support notes and error logs so that they are objectively far more knowledgeable than the average human call center assistant. Plus, they don’t get tired. As exhausting as you may find it to be on the phone with customer service for 15 minutes, imagine how the person on the other end of the line—who may be nine hours into her shift—feels! Most of us have probably already worked with text-based customer-service bots on the help sections of many web pages, perhaps without even knowing it. If the bots can objectively outperform most humans, most of us will probably soon reconcile ourselves to machine-based customer service.

Indeed, while pessimists have worried about the massive job losses that the AI revolution is likely to impose on many industries, optimists have argued that the jobs lost will often be the ones that almost no one would want to do anyway—the dull, dehumanizing jobs in which the workers were already reduced to little more than robots. If the jobs aren’t fit for humans anyway, why not embrace that fact and free up human workers for more fulfilling work?

There’s truth in this argument, but we should recognize what a dim picture this paints of economic and technological progress. Once upon a time, customer service meant going to the craftsman who made the product and asking him why it wasn’t working. Industrial development and mass production put an end to this, separating those who made gadgets from those who fixed gadgets or answered questions about them.

In other words, only once technology itself has fully dehumanized customer service, rendering it miserable for those on both ends of the call, does technology then propose itself as the solution to the problem it has caused. Indeed, this is a standard feature of many modern technologies. Food processing made us fat, so we needed to invent Ozempic to make us thin again. Cars made us live far apart, so we needed Zoom to reconnect us. This isn’t always bad, of course. Sometimes a new technology really does definitively solve the problems that a previous technology had created. For instance, refrigeration technology initially released chlorofluorocarbons into the atmosphere, creating the famous “ozone hole.” Subsequent technological developments, though, enabled us to have our cake and eat it, too, offering safe, sustainable refrigeration.

Sometimes, though, the technological “solution” just continues a feedback loop. Zoom helps us feel more connected when we’re apart, but, at the same time, it makes us more likely to stay apart, compounding the problem it was designed to “solve.”

Which sort of solution will AI customer service provide? Will it put an end to a dehumanizing class of work and free up humans for more meaningful interaction? Or will it continue to rewire our expectations of what counts as “service” so that we come to prefer bots and see human beings as dim-witted, ill-mannered inconveniences that we’d prefer to keep out of sight and out of mind? If the latter, it is hard to see how the new jobs created by this industry disruption will prove more “liberating” or “humanizing” since we are only free to be authentically human when we are engaging with one another.

AI can be deployed to repair—or to exacerbate—the dehumanizing effects of the industrial age. Only a civilization that still understands what it means to be human, however, has a decent shot of using these technologies to help rather than to harm.