AI-generated faces are growing increasingly popular in advertising. But the models are flawed: They often perpetuate outdated beauty standards and make mistakes, such as misplaced teeth and misshapen ears

FACE-GENERATING AI SYSTEMS ARE BECOMING INCREASINGLY POPULAR in the advertising industry, with companies like DeviantArt, Prisma Labs, and Meta using the systems to create futuristic, photorealistic images of people. The AI algorithms, which are typically trained on large datasets of images, can create diverse images of people, including different ages, races, hairstyles, and facial expressions. This allows advertisers to diversify their campaigns without having to shoot images of real people.<br><br>But while these AI systems may present an innovative solution for the advertising industry, they have limitations. The algorithms can make mistakes, such as creating images of celebrities or public figures not related to the brand. For example, a recent Prisma Labs campaign created an image of the comedian Ali Wong. This can lead to confusion and misidentification.<br><br>Furthermore, the algorithms may perpetuate outdated beauty standards, such as idealized facial features and hairstyles. This can contribute to the commercialization of beauty and reinforce harmful beauty standards. Additionally, the algorithms can create stereotypical or racist images, which can be offensive and damaging.<br><br>Therefore, while AI-generated faces may present a convenient solution for the advertising industry, they also raise ethical concerns. Advertisers should approach these systems with caution and consider the potential consequences of using them.