AI, Anti-Intellectualism, and the Era of American Illiteracy

Photo by Compare Fibre on Unsplash

As an artificial intelligence trainer, I often receive various projects to choose from when working with AI. While I enjoy projects that involve prompt engineering, exploring critical theory, or correcting code, I have always disliked projects that require me to correct the AI on tasks it cannot perform.

For example, did you know that artificial intelligence cannot see pictures? While this might sound obvious, as much as people love to talk about computer vision, computers don’t have eyes. When artificial intelligence interprets an image, it doesn’t give information based on what the photo actually looks like but uses pattern recognition and image processing to tell you what might be in the image. This means that training a model to improve its photo recognition ability can often feel like an uphill battle.

From English to AI: How Writers are Crafting the Future of LLMs

Now that you know AI can’t see, did you know that large language models lack reading comprehension skills? Now, this statement might be a little more surprising. How does it work if it’s not comprehending what it’s reading? Well, that is the very…