Many of us have seen the breakthroughs in AI-generated art. From a simple text prompt, AI engines like MidJourney, Stable Diffusion and DALL-E can generate entire scenes and videos. Soon, we’ll likely have entire procedurally generated 3D metaverses. Unfortunately, the biases that have been intertwined in AI for some time now are more visible than ever.
AI image generation works in a similar pattern to most common AI: A team assembles a large amount of given set of data (statistics, text, speech, images, video) and then spends effort to “train” the AI by having a team go through and identify the target objects in each image or statistically analyzing which words usually seem to be positioned next to each other.
For generating art, AI engines have been fed thousands of paintings, drawings and photographs. These images all were tagged, sometimes manually but increasingly automatically, as the AI recognized objects. These recognitions act cumulatively, so now an AI “knows” what a cat looks like, “knows” what a whiteboard looks like, and “knows” the visual patterns that are unique to Matisse, Dali or Banksy.
These patterns, and the data sets from which they derived, are the problem. For decades, the cultural pattern for “developers” was mostly white males. So, prompting an art-generating AI like MidJourney with words like “tech developers,” “laptops” and “business” will generate images with men, not women.
AI plays to stereotypes. Unless developers are very specific, the AI will produce images that will only reinforce the stereotypes.
Specificity is required to reduce bias in AI. The prompt will do what it’s told, so it is up to the person entering the prompt to break the stereotypes by typing in “female tech developers,” “laptops” and “business.”
Biases in AI can be overcome with forethought and conscious actions. At a programmatic level, it is key for teams building AI engines to identify and understand how biases creep into the AI, both from a cultural stereotyping perspective as well as the mathematical set of statistics upon which the AI draws its output.
The AI will improve in accuracy and creative ability. As AI continues to march into more areas of our lives, we will see an increasingly important role for “AI ethicist,” “AI bias manager” and “AI culturalist.” With effort, AI bias can be counteracted.
Dave Jenkins is VP of technology curation at Iterate.ai
