Subtlefakes: New AI-Generated Nonconsensual Images on X Are More Convincing

Evolution of AI-Generated Nonconsensual Content

The landscape of AI-generated nonconsensual imagery has evolved significantly since the emergence of deepfakes in 2017. Early deepfakes, while technologically advanced for their time, were often clearly identifiable as fabricated, such as face-swapped identities in pornographic videos. The primary use of this technology has consistently been for harassment and degradation, particularly targeting women.

Introduction of "Subtlefakes"

A new category of AI-generated nonconsensual images, dubbed "subtlefakes," has begun circulating on X. These images differ from traditional deepfakes by featuring subtle alterations to celebrity photos, making them appear more revealing or provocative rather than outright fabricating explicit content. This subtlety makes them more convincing and harder for viewers to immediately discern as fake.

Increased Sophistication in Malicious Use

The rise of subtlefakes indicates an advancement not only in AI technology but also in the methods employed by those who use it maliciously. Similar to "cheap fakes" which manipulate truth without advanced AI, subtlefakes represent a more sophisticated form of synthetic media. This development poses a greater challenge as the images are designed to be more believable, potentially fooling a larger audience.

Impact and Implications

Nonconsensual AI-generated media, regardless of its believability, causes psychological and material damage to its subjects. The increased convincingness of subtlefakes amplifies this harm, as the line between reality and fabrication becomes blurrier. This trend highlights the ongoing struggle for internet platforms and legal frameworks to keep pace with rapidly advancing AI capabilities and their misuse.