A parent used AI image generators with their kids and noticed the tools default to narrow beauty standards: blonde hair, brown eyes, olive skin. By using thumbs-up and thumbs-down feedback on these platforms, users can help train models toward more diverse representations.
One of my most, uh, the most fun that I have, especially with my kids, is becoming a producer rather than a consumer. Mm-hmm. Like using AI to build images. Mm-hmm. When you realize how much clarity of thought you need to have when you're building imagery, you'll realize, oh, this model is really working with very little information about what beauty is. As a matter of fact, uh, part of the study that I read, they said that if you ask a model that's not been trained on what beauty is, it'll say blonde hair, brown eyes, and olive skin. And that, of course, is a very small representation of the world. So in that world, how do we actually train the model to be better? So in every model, there's a thumbs up and a thumbs down on all of these generative tools. And so I encourage everyone from companies to individuals to just take the opportunity to tell a model, nope, you didn't quite get it. And I think this proactive approach to building AI, it's new for us. Of course, we're used to just sitting back and absorbing.
Take the Unlocking AI Challenge and Get the Keys to Unlocking Time and Revenue for Your Business
Don't Get Left Behind: The AI Executive Challenge #ai #business #fearofmissingout
