Is AI-Generated OK?

Daniel Tunkelang
3 min readMay 8, 2023

For as long as we have had needs, long before servers started asking us “Sorry, we don’t have Coke, is Pepsi ok?”, we have accepted substitutes for the products and services we seek. Indeed, fungibility is so universal that non-fungibility, digital or otherwise, is the exception.

Today, however, there is a groundswell of concern that generative AI will replace human creativity and even humanity in general. Creators across a variety of media fear being copied and replaced. Workers fear losing their jobs and livelihoods to automation. Instead of crying that the Skynet is falling, doomsayers now forecast the commodification of humanity.

The doomsayers have a point. Automation has a long history of commodifying human labor, as evidenced by the growing number of words like “calculator” and “computer” that once described people and now define the machines that have automated their jobs. Over time, the number of things we believe that only humans can do has dwindled in the face of technological advancement. If we are to cling to human exceptionalism in the face of automation, we have to come to terms with its dwindling scope.

That said, there have always been such doomsayers, most famously the 19th-century Luddites who raged against the Jacquard loom. Today, their concern about the automation of textile manufacturing seems quaint. Nonetheless, every generation has produced its version of Luddites, fearing and resisting new technological advances. Yet every generation has also embraced these advances, and the world has not ended.

Is generative AI different? Will the availability of AI-generated text, imagery, video, and who knows what else produce substitutes that people accept wholesale in the place of human-generated creations? If so, are we witnessing the swan song of human creativity?

Personally, I don’t think so.

It is not that I see inherent limitations on the power of generative AI. Today’s technology has already surpassed my expectations, and I expect it to keep improving in quality and scope. I am skeptical of most statements that begin with “AI cannot” — unless they are qualified with the word “yet”.

Rather, I believe that people place a particular value on human connection, even when that value may seem sentimental or even irrational. For example, we appreciate handwritten notes, even though we know that machines can produce better penmanship. We respond positively to active and reflective listening techniques that are highly amenable to automation, but I strongly suspect that people are far less eager to be “heard” by machines than by fellow human beings. Many of our interactions rely on an unstated assumption of human connection.

Of course, that is not true of all of our interactions. Vending machines and ecommerce sites prove that we are quite willing to shop without personally interacting with shopkeepers. Search engines and other information access tools have reduced our need to interact with librarians, scholars, and other experts. In these and many other cases, we appreciate the efficiency and convenience of automation, even if it means forgoing human connection.

Will we feel that way about art, music, and television? It is hard to say. We are in the early days of generative AI, and I think it will take some time for us to converge to new cultural norms.

But I am skeptical that we will turn to AI for friendship, companionship, or other needs that come down to human connection. It is not that generative AI cannot produce the right words. Perhaps it can, especially when it is trained on our increasingly digital behavior.

But I do not believe that human connection is an area where we will be willing to accept machines as substitutes.

Sometimes, AI-generated is not OK.