With the World Diet and Hydration Week 2024 beginning at the moment, researchers have introduced an intriguing discovery — shoppers usually favor AI-generated photos of meals over actual meals photos, particularly when they’re unaware of their true nature. The brand new findings have been printed in Meals High quality and Choice.

In response to the researchers, the outcomes recommend that AI-generated meals visuals excel at enhancing the attraction of depicted meals by leveraging key options similar to symmetry, form, glossiness, and general lighting and color. All of those are recognized to contribute considerably to the attractiveness of meals imagery.

Even refined tweaks in positioning could improve the attraction of AI-generated meals photos. Lead writer Giovanbattista Califano (Division of Agricultural Sciences, College of Naples Federico II) defined: ‘As people, we are likely to really feel uneasy with objects pointing in the direction of us, decoding them as threats, even when it is simply meals. When tasked with replicating meals pictures that includes gadgets pointing on the viewer, similar to a bunch of carrots or a chunk of cake, the AI usually positions the meals in order that it would not straight level on the viewer. This warrants additional research, nevertheless it’s believable that this strategy enhances the perceived attractiveness of the depicted meals.’

Within the examine, the researchers requested 297 contributors to price actual or AI-generated meals photos on a scale from “Under no circumstances appetizing” to “Extraordinarily appetizing.” The photographs depicted a variety of pure, processed, and ultra-processed meals, from apples and carrots to chocolate milkshakes and potato fries. When contributors have been informed how every picture had been created — whether or not by images or AI — they tended to price actual and AI-generated variations equally interesting. Nonetheless, when contributors have been unaware of the picture creation course of, the AI-generated model was persistently rated as considerably extra appetizing than the true meals picture.

Research supervisor and co-author Professor Charles Spence (Division of Experimental Psychology, College of Oxford) mentioned: ‘Whereas AI-generated visuals could supply cost-saving alternatives for entrepreneurs and the trade by lowering the price of commissioning meals photoshoots, these findings spotlight potential dangers related to exacerbating ‘visible starvation’ amongst shoppers — the phenomenon the place viewing photos of meals triggers urge for food and cravings. This might probably affect unhealthy consuming behaviours or foster unrealistic expectations about meals amongst shoppers.’

Moreover, the researchers additionally discovered that AI-generated photos are likely to depict meals to seem extra energy-dense in comparison with the originals, notably within the abundance portrayed. As an illustration, AI could improve the variety of fries within the picture or add extra whipped cream to a dessert. Provided that people have an evolutionary drive to pay extra consideration to energy-dense meals, this raises issues that widespread dissemination of such idealized meals photos may promote cue-induced consuming of unhealthy meals.

Moreover, with the worldwide motion in the direction of extra sustainable consumption patterns, together with the promotion of ‘ugly’ fruit and veggies, there’s a concern that fixed manufacturing of AI-enhanced meals photos may nudge shoppers in the direction of an unrealistic normal of how pure meals ought to look, probably harming sustainability efforts.

LEAVE A REPLY

Please enter your comment!
Please enter your name here