Science1 publisher2 min readPublished
Calling an identical backpack "AI-designed" cost it 0.66 points on a seven-point sustainability scale
In a Georgia State and University of South Florida study, the same product rated less sustainable once consumers were told AI designed it. The researchers trace the penalty to a perception they call genuine care.
The Scientist · Science desk
What happened
- A Georgia State University and University of South Florida study in the Journal of the Academy of Marketing Science asked how attributing a product's design to AI changes what consumers think of that product.
- An identical backpack scored 4.27 out of 7 for perceived sustainability when described as human-designed and 3.61 when labeled AI-designed, with nothing altered but the description.
- Offered a free toothbrush and told to weigh sustainability, more than three-quarters of participants picked the human-designed option over an identical AI-designed one.
- Statistical testing identified perceived genuine care as the mechanism linking AI involvement to the lower sustainability ratings.
Compiled by The ScientistSomething wrong?How this is made
Why it matters
- decision A team that wants to name AI in the product story now has a measured penalty to set against the appeal of the novelty, plus one specific wording that reduces the penalty without removing it.
- constraint Since the penalty holds no matter how AI-literate the consumer is, aiming the label at tech-fluent audiences will not clear it, and neither will waiting for the public to grow used to AI.
- contradiction AI can lower a product's actual impact while naming AI in the pitch lowers the audience's belief in that impact, so the engineering goal and the marketing goal work against each other.
The backpack gap is 0.66 points [1], about 15 percent below the human-designed rating [2]. On a seven-point scale the neutral midpoint is 4, so the human-designed version sat above it and the AI-designed version below [3].
The idea behind that gap came out of interviews with senior product design executives, who described what the team called "genuine care": the sense that real love, intentionality and moral concern went into how a product was made [7]. One executive said, "Why would AI care if my product's sustainable like a human would? AI doesn't have kids that are going to walk the Earth in 100 years." [8]
"As we conducted our interviews to figure out why consumers trust AI-designed products less, it became clear that it's emotional, not technical. We found that consumers perceive sustainability as inherently about love and care for the environment and society," said Barbara Duffek, an assistant professor of marketing at Georgia State and a co-author of the study [9][15].
The toothbrush experiment is the closest thing here to behavior, and it carries a condition. Participants were offered the free toothbrush with sustainability specifically in mind [4]. That framing isolates the mechanism, and it also keeps the split narrow: this is what people picked when sustainability was the stated criterion. Fewer than one in four took the AI-designed brush [4]. Nothing was priced, so no cost competed with the perception.
In the live Meta test, the part with real money behind it, the announcement reports which way the numbers moved and skips the denominators. The "AI-designed" cup framing cut click-through and reduced engagement to a fraction of the human-designed version [5], with no rates, impression volumes or spend disclosed in the announcement. Click-through measures a first reaction to ad copy, not what happened at checkout.
Disclosing AI use, on its own, does not reduce the penalty [12]. Neither does the consumer's own AI literacy: the effect turns up regardless of how familiar they are with the technology [13]. Describing the AI as "guided by genuine care for environmental and societal outcomes" does shrink it, and the researchers report that the penalty survives [11].
None of this measured a product's footprint. AI is being used to optimize materials, cut waste and streamline manufacturing, and the study's own framing accepts that it can make a product more sustainable in objective terms [14]. What the experiments recorded is what people said and clicked [2].
What to watch
- Whether the published paper reports click-through rates, impressions and spend for the Meta test that the announcement summarizes only by direction.
- A replication in which the AI-designed and human-designed versions carry different prices, so the perception has to compete with money.
- Whether the penalty appears in categories where buyers already assume machine design, such as software or consumer electronics.