AI Political Images Are Flooding the Feed — and Losing Their Force
A new monthly column on technology and contemporary art argues that the internet has made political imagery faster to produce, but far easier to forget. In the latest installment of “Press any key to continue,” visual artist, researcher, and Contemporary Istanbul Foundation artistic director Sarp Kerem Yavuz examines how AI-generated images and platform distribution are reshaping dissent, propaganda, and the visual language of crisis.
Yavuz describes the current wave of machine-made political pictures as “AI slop” — a shorthand for images that can be timely, even viral, without carrying much lasting weight. He points to AI-generated images of Donald Trump as Jesus, which drew criticism from both MAGA supporters and the broader public, as well as animated Lego-style videos featuring militarised IRGC minifigures rapping diss tracks about US and Israeli attacks in the Middle East. The videos, produced by Explosive Media, were deemed dangerous enough for YouTube to suspend the account for “violent content.” Representatives for Explosive Media said the group is not affiliated with the Iranian government, though the regime is a customer.
For Yavuz, the problem is not only the imagery itself, but the system that circulates it. Social platforms reward speed, repetition, and legibility, often flattening political art into content optimized for attention rather than endurance. That concern leads him back to the Gezi Park protests in Istanbul in 2013, where humor, language, and visual wit became tools of resistance against state repression. The protests, he writes, shaped an entire generation of Turkish artists, including himself.
The comparison is telling. At Gezi, dissent was inventive, local, and difficult to translate without losing its charge. Today’s AI-generated political visuals can be instantly distributed across Instagram, YouTube, and other platforms, yet their very ease can drain them of specificity. Yavuz’s column suggests that the real question is no longer whether technology can produce political images, but whether those images can still carry memory, critique, or consequence once they enter the algorithmic feed.
