Exploring the Challenges AI Faces in Accurately Generating Images of Kamala Harris

When Elon Musk shared an image on X last week showing Kamala Harris dressed as a “communist dictator,” it was clearly a fabrication since Harris is not a communist nor known for Soviet cosplay. Moreover, the woman in the photo, which appears to have been created by X’s Grok tool, only slightly resembled the Vice President. Observers quickly pointed out the discrepancies.

“AI still is unable to accurately depict Kamala Harris,” one X user commented, noting that the image posted looked more like “some random Latina woman.”

“Grok put old Eva Longoria in a snazzy outfit and called it a day,” another remarked, highlighting the resemblance of the “dictator” in the image to the actress known from Desperate Housewives.

“AI just CANNOT replicate Kamala Harris,” another user asserted, criticizing the algorithm’s failure to accurately represent an American of South Indian and Jamaican heritage.

Many AI images of Harris are similarly bad. Meanwhile, a tweet featuring an AI-generated video showing Harris and Donald Trump in a romantic relationship—it culminates in her holding their love child, which looks like Trump—has nearly 28 million views on X. Throughout the montage, Harris morphs into what look like different people, while the notably better Trump imagery remains fairly consistent.

When we tried using Grok to create a photo of Harris and Trump putting their differences aside to read a copy of WIRED, the results repeatedly depicted the ex-president accurately while getting the vice president wrong. Harris appeared with varying features, hairstyles, and skin tones. On a few occasions, she looked more like former First Lady Michelle Obama.

Grok is different from some high-profile AI image generators in that it allows users to create faked photos of political figures. Earlier this year, Midjourney began blocking its users from creating images of Trump and President Joe Biden. (The ban extends to Harris.) The move followed publication of a report by the Center for Countering Digital Hate that found that the tool could be used to generate a range of politically charged images.

Similarly, OpenAI’s ChatGPT and Google’s Gemini refused to produce images of Harris or Trump in WIRED’s testing. Meanwhile, a number of open source image generators will, like Grok, produce images of politicians. WIRED found one such model, Stable Diffusion, also produced not-great pictures of Harris.

Modern AI image generators use diffusion models to create images from textual prompts. These models are trained using massive sets of labeled images, often sourced from the internet. Joaquin Cuenca Abela, CEO of Freepik, explains to WIRED that these generators struggle more with depictions of Harris than Trump due to a smaller dataset of accurately labeled images of her.

Despite her prominence, Harris has been less frequently photographed than Trump, a fact supported by a search on Getty Images, which shows significantly fewer images available for her. According to Cuenca Abela, Harris, being relatively new to the presidential scene, is considered “a new celebrity” by AI image creators, and it usually takes a few months for the AIs to catch up.

The fact that Harris is a Black woman of Jamaican and Indian lineage also plays a role, according to Irene Solaiman, head of global policy at Hugging Face. She points out that poorer facial recognition capabilities for darker-skinned individuals and feminine features could impair the effective sorting and labeling of images. This issue connects back to the 2018 Gender Shades study, which was led by MIT researcher Joy Boulamwini and Timnit Gebru, who later founded the Distributed Artificial Intelligence Research Institute.

Furthermore, the quality of AI portrayals of Harris might also be influenced by the intent behind the image creation. Images are not always designed to be photorealistic but rather to support specific narratives, as noted by Hany Farid, a specialist in deepfake detection and co-founder of GetReal Labs.

In other words, those sharing AI-generated images of Harris may often be more interested in producing meme-worthy scenarios than refining the realism of her likeness. The “communist dictator” image shared by Musk and the video in which Harris holds her Trumpy baby both serve to ridicule and denigrate the Democratic candidate rather than spread disinformation.

Ari Lightman, professor of digital media and marketing at Carnegie Mellon University’s Heinz College, says some people may even be selecting bad Harris images on purpose in an effort to emphasize the idea that she is a fraud. “This is an AI-generated communications era,” Lightman says. “If it’s done crudely, it’s designed to send a message.”

Total
0
Shares
Leave a Reply

Your email address will not be published. Required fields are marked *

Previous Article

NBA 2K25 Review: A Stellar Game Overshadowed by Pay-to-Win Concerns

Next Article

AMD Shifts Focus: Embracing AI Over Traditional Gaming

Related Posts