As AI Expands, Consumers Will Still Prefer Some Products Made With a Human Touch

Our research suggests an answer that is surprisingly simple, but one that may prove powerful in many industries: In many cases, consumers prefer products and services that rely on human labor (vs. robots or AI).

In our work, my colleagues and I focus on “symbolic” consumption, which is motivated at least in part by the desire to be a particular type of person—related to group membership, personal values, interests, status or style. For example, a consumer may purchase a glass vase to display aesthetic taste, wealth or group membership (e.g., a vase with college insignia). Other times, by contrast, consumption is motivated by more practical, instrumental reasons, such as buying a glass container simply to store food.

Reading glasses

In a series of studies, we found a stronger preference for human labor when consumption has symbolic qualities. Human labor is associated with uniqueness, which consumers seek in symbolic consumption, while machines are linked to repetition and standardization.

With Armin Granulo from the Technical University of Munich and Christoph Fuchs from the University of Vienna, we asked people to read consumption scenarios and randomly varied the type of labor (human vs. machine) and motivation (more vs. less symbolic) described. We then measured how consumers chose products or their purchase intentions.

In one study, consumers preferred reading glasses where the frame—a fashion accessory, and hence symbolic—was made by hand versus machine. But they preferred glasses with machine-made lenses since they are less symbolic, and more purely functional.

In another study, we asked participants to imagine going to a tattoo parlor employing both human and robotic service providers, and we varied the visit’s purpose. At times, we said the visit was to get a tattoo, which is perceived to be highly symbolic, and at other times we said it was to remove a tattoo, which is perceived to be more practical and relatively less symbolic. People preferred human labor in general, but this preference was much stronger when respondents were considering getting a tattoo versus removing one.

In other studies, we asked participants to consider purchasing the same product for different reasons. First, we asked them to imagine being a doctor buying the print of a skull for the office. Half of the participants were told they were buying the poster to decorate the office (a symbolic motive), while the other half that they were buying the poster to illustrate anatomy to patients (an instrumental motive). We then asked participants to choose between two similar posters and randomly varied whether one of them was described as drawn by a human or by an algorithm. Participants more often preferred the human-drawn poster when they were buying it to decorate the office than when they were buying it to illustrate anatomy.

Several of our studies confirmed that the reason consumers prefer human labor more in symbolic contexts is because they link human labor to greater perceived uniqueness. In one of the studies where people chose a poster to decorate a doctor’s office, the relative preference for human labor was particularly strong among individuals who placed more importance on being unique.

Fading preference?

As algorithms and robots continue spreading in the economy and consumers get used to automation in front-line services, the general preference for human labor we tend to see in many contexts today may weaken or even disappear.

However, our studies used a variety of contexts, including some where production has been relying heavily on machines for decades (reading glasses) and some where the use of machines is much more recent (tattoos). Our effects emerged in both, so we don’t believe the effects are likely to disappear soon.

In fact, it’s possible the effects will grow stronger with time. As AI and robots diffuse through the economy, we will grow more comfortable and familiar with them, and this will likely result in greater consumer acceptance. At the same time, the special value of human labor may also increase with the same trend, as human labor is no longer the expected default and hence could contribute even more to feelings of uniqueness.

We know the AI revolution will require large adjustments in labor markets. Many new jobs will be created, many lost. But regardless of how things eventually pan out, how can companies justify keeping existing workers meanwhile?

Firms tend to make automation decisions based on internal factors, such as operational efficiency. However, companies should also consider demand-side factors, such as the preference for human labor we’ve found in our research. This way, they can identify the business case for human employment, even when supply-side factors may speak against it. Companies should shift the conversation from what can be automated to what should be automated.

The key may be companies’ ability to let consumers know when products are made by humans—from packaging or advertising, for instance, or in-store displays.

Perhaps standards and certifications, like for Fair Trade products today, will emerge to help companies capitalize on the distinct value of human labor. In a few years, people may seek products and services that involve human labor, at least when these speak to consumers’ sense of who they are. Machines are increasingly able to do what we do, cheaper and often better. But only people offer companies the opportunity to entice consumers with products “Made by Humans.”

The value of human labor in the age of AI? Consumer demand may turn out to be the prosaic answer.

Stefano Puntoni is the Sebastian S. Kresge professor of marketing at the Wharton School, University of Pennsylvania, and the co-director of the Wharton Impact of Technology Initiative. He can be reached at [email protected].

Tap to View