Artificial Images in Global Health

The increasing quality of photorealistic generative artificial intelligence (AI) tools are making the detection of fake  pictures more difficult. In this era of disinformation, misinformation and fake news, it falls to the consumer to  determine what is true. Here are some tips and prompts to help you become engaged and emancipated spectators. 

AI can be used to raise awareness of important issues without burdening populations that are overrepresented in aid campaigns. But in this case, it appears the choice was to reproduce a crass stereotype. What can we determine about software bias when the only difference in the prompts for the middle and right-most images is the word “Africa”?

AI can be used to imagine futures, particularly those we don’t want. In this case, AI Samurai created a series of images to imagine what American elementary schools may look like in 2030 should gun control measures not be enhanced. Powerful and potentially beneficial as this may be, no technical safeguards are in place to ensure people don’t fall for the fakes.

AI can be used to imagine current realities, particularly those realities experienced by people with few or limited rights. Described as ‘post-photography’, 90-Miles is an AI reporting experiment by photographer Micheal Christopher Brown. To publish photographs of Cubans escaping to Florida would put them at risk. Brown turned to AI to created safer images based on his professional photojournalistic pictures. Truthfulness over accuracy?

As intelligent as AI developers claim it to be, and as much promise the software holds, it is still in its infancy like the camera, the radio and the television once were. Media scholars will remind us that technology is always a human creation. Humans are as creative as they are diverse. As such, AI’s fate is not sealed. Ask yourself: What parts of our humanity do we want AI to reflect? What parts of our humanity do we want to amplify?