Artificial intelligence can help designers optimise materials, reduce waste and explore large numbers of product configurations. But a new marketing study suggests that technical capability is only part of what consumers see when they judge whether a product is sustainable. The identity of the designer can change the judgement even when the product itself does not.
Researchers Barbara Duffek and Dipayan Biswas found that products described as designed by artificial intelligence were consistently perceived as less sustainable than equivalent products attributed to human designers. Across a programme of experiments, interviews and a live advertising test, the researchers traced much of this penalty to what they call genuine care: the belief that a designer put emotional concern, moral intention and care for environmental and societal outcomes into the product.
The findings are important because companies are increasingly using AI during product development while simultaneously trying to communicate stronger sustainability credentials. The study suggests those two messages can collide. Consumers may recognise AI as efficient or technically capable while still doubting whether a machine can possess the kind of care they associate with sustainable design.
A sustainability judgement changed when only the designer changed
The research, published in the Journal of the Academy of Marketing Science, combined five preregistered experiments and a Meta A/B field test in the main analysis, with six additional supporting studies reported in the web appendix. The researchers also conducted qualitative interviews with corporate product-design executives to understand how professionals conceptualise human and AI involvement in design.
One of the clearest demonstrations involved 193 participants recruited through Prolific. Participants encountered a naturalistic social-media-style advertisement for a backpack. The underlying product was the same, but the stated designer differed. On a seven-point measure of perceived sustainability, the AI-designed backpack received an average score of 3.61, compared with 4.27 when the backpack was described as human-designed.
The difference was statistically significant, F(1, 191) = 14.31, p < .001, with an eta-squared effect size of 0.070. This matters because the manipulation did not establish that the AI-designed version used worse materials, consumed more energy or produced more waste. Instead, changing the attributed design agent was enough to alter sustainability perceptions.
A supporting replication explicitly gave the human designer a name rather than referring generically to a product designer, and the same broad pattern emerged. Another study included a condition in which no designer information was supplied, helping the researchers examine whether consumers were rewarding human design, penalising AI design, or both.
The effect extended beyond survey ratings
Perceptions are commercially important when they translate into behaviour. The researchers therefore tested the effect outside a conventional survey by running a live Meta advertising experiment for a reusable cup. Both advertisements promoted the product in a sustainability context, but one described it as AI-designed and the other as human-designed.
Across the campaign, the advertisements reached 11,440 unique users and generated 15,884 impressions. The AI-designed condition reached 3,968 users and received 77 clicks, while the human-designed condition reached 7,472 users and received 158 clicks. Based on impressions, the click-through rate was 1.24% for the AI condition and 1.63% for the human condition. The difference reached statistical significance, z = 1.98, p = .048.
The field result does not mean that every disclosure of AI involvement will reduce advertising performance. A single campaign cannot reproduce the full complexity of real brands, established reputations, price differences or consumers’ prior knowledge. It does, however, strengthen the central finding by showing that the design label can influence observable behaviour rather than only responses to questionnaire items.
Why AI lost the sustainability advantage
The researchers’ interviews with product-design executives pointed toward a distinction between technical efficiency and emotional intentionality. Human design was commonly associated with craftsmanship, moral concern and an ability to care about the people and environment affected by a product. AI was more readily associated with speed, precision and data-driven efficiency.
That qualitative pattern led to the concept of genuine care. The researchers then tested whether this perception statistically explained the sustainability gap.
In one experiment, participants again rated an AI-designed product as less sustainable than a human-designed product. Average sustainability ratings were 3.78 for the AI condition and 4.54 for the human condition, F(1, 197) = 18.22, p < .001. Participants also perceived substantially less genuine care when AI was identified as the designer. The coefficient for the effect of AI attribution on genuine care was b = -1.73, SE = 0.20, t = -8.83, p < .001.
Genuine care, in turn, positively predicted perceived sustainability, b = 0.51, SE = 0.05, t = 9.48, p < .001. A mediation analysis using 5,000 bootstrap resamples produced an indirect effect of -0.88, with a 95% bootstrap confidence interval from -1.19 to -0.61. Because that interval did not include zero, the analysis supported genuine care as an important pathway linking the stated designer to sustainability perceptions.
Additional analyses tested alternative explanations. According to the study, genuine care remained the significant mediator when cognitive intentionality and perceived effort were considered alongside it. This suggests that consumers were not simply assuming that humans worked harder or thought more deliberately. The emotional meaning attached to care was particularly important.
Communicating care narrowed the gap
The researchers next asked whether the disadvantage was fixed. In another experiment, they manipulated whether participants received an explicit cue that the designer had been guided by genuine care for environmental and societal outcomes.
Without that cue, perceived sustainability averaged 3.95 for AI-designed products and 5.11 for human-designed products. When genuine care was explicitly communicated, the AI average rose to 4.94 while the human-designed product averaged 5.48. The interaction between design agent and the care cue was statistically significant, F(1, 386) = 6.28, p = .013.
The care message therefore narrowed the AI sustainability penalty but did not eliminate it. For AI-designed products specifically, the increase from 3.95 to 4.94 was significant, F(1, 386) = 32.11, p < .001. The corresponding increase for human-designed products was much smaller, from 5.11 to 5.48.
This distinction offers a practical lesson for brands. Simply telling consumers that AI was used responsibly may not communicate the same thing as showing the values and environmental goals that guided its use. Supporting studies reported by the authors found that transparency cues alone did not remove the effect, and AI literacy did not meaningfully moderate it. Familiarity with AI therefore did not appear sufficient to erase the human-design advantage.
What the findings mean for companies using AI
The study exposes a potential communication problem for businesses. AI can be useful precisely because it can optimise complicated design objectives, potentially including material efficiency, durability or waste reduction. Yet if consumers use the identity of the designer as a shortcut for judging sustainability, objective improvements may not automatically produce stronger sustainability perceptions.
For marketing teams, this means that an AI disclosure can carry symbolic information beyond the factual statement that a tool was used. Consumers may interpret the disclosure as evidence about intention and values. A company that can demonstrate why an AI system was directed toward environmental goals, how human values shaped its objectives, and what measurable sustainability outcomes resulted may therefore have a stronger message than a company that foregrounds technological novelty alone.
The results also caution against the opposite conclusion. They do not show that human-designed products are objectively more sustainable. The researchers explicitly distinguish perceived sustainability from measured environmental performance. A human designer can create an environmentally damaging product, and an AI system can contribute to a design that reduces material or energy use. The experiments reveal how consumers interpret design agency, not which type of designer produces the smallest environmental footprint.
Important limitations remain
Several limitations shape how the findings should be applied. Many of the studies used controlled scenarios in which the designer label was deliberately made salient. Real purchasing decisions involve price, brand reputation, product quality, certifications, environmental claims and previous experience, all of which may strengthen or weaken the effect.
The live advertising test adds behavioural evidence, but it represents one product context and a relatively small advertising campaign. Future research could test whether the same pattern persists for expensive durable goods, services, industrial products or categories in which algorithmic optimisation is already expected.
The concept of AI design is also changing quickly. Consumers may react differently to a product created autonomously by an AI system, a product developed by a human using AI as one tool, and a product created through sustained human-AI collaboration. As those distinctions become more familiar, the meaning consumers attach to an AI-designed label could change.
For now, the evidence suggests a clear psychological tension. Consumers can recognise the technical promise of artificial intelligence while still treating sustainability as a deeply human signal. When environmental responsibility is interpreted partly through care, intention and moral concern, efficiency alone may not be enough to convince them.
Source Information
Study: Duffek, B. and Biswas, D. (2026). AI and human designers: How consumers see the genuine care in product design as more sustainable.
Journal: Journal of the Academy of Marketing Science
DOI: 10.1007/s11747-026-01170-4
Publication date: 6 July 2026
Article type: Original empirical research, open access
