Human rights advocacy group Amnesty International has retracted images generated by artificial intelligence that it used in a campaign to publicize police brutality in Colombia during national protests in 2021.

The group was criticized for using AI to produce the images for its social media accounts, according to reports. One image, in particular, was highlighted by The Guardian on May 2.

It depicts a woman being dragged away by police during Colombia’s protests against deep and long-standing economic and social inequalities in 2021.

However, a closer look shows a few discrepancies in the image, such as the uncanny-looking faces, dated police uniforms and a protestor who appears to be somehow wrapped in a flag that is not the correct flag of Colombia.

The bottom of each image also carries a disclaimer saying the images are produced by an AI.

AI-generated image from Amnesty International. Source: Twitter

Amnesty International told The Guardian it chose to use AI to generate images to protect protesters from possible state retribution. Erika Guevara Rosas, director for Americas at Amnesty, said:

“We have removed the images from social media posts, as we don’t want the criticism for the use of AI-generated images to distract from the core message in support of the victims and their calls for justice in Colombia,”

Photojournalists criticized the use of the images, commenting that in today’s highly polarized era of fake news, people are more likely to question the media's credibility.

AI-generated image from Amnesty International. Source: Twitter

Media scholar Roland Meyer commented on the deleted images, stating, “Image synthesis reproduces and reinforces visual stereotypes almost by default.” Meyer added that they were “ultimately nothing more than propaganda.”

Other images, now deleted by Amnesty, were shared by Twitter users in late April.

AI-generated image from Amnesty International. Source: Twitter

Related: Here’s how the crypto industry is using artificial intelligence

AI is being increasingly used to generate images and visual media. In late April, HustleGPT founder Dave Craige posted a video of the United States Republican Party using AI imagery in its political campaign.

“We all knew that AI and deep-fake images were going to make it to politics, I just didn’t realize it would happen so quickly,” he exclaimed.

Cointelegraph contacted Amnesty for comment but had not received a response at the time of publication.

Magazine: AI Eye: ‘Biggest ever’ leap in AI, cool new tools, AIs are real DAOs