Great post, Chris. I can think of a couple additional elements of critical AI literacy that would be helpful within your definition of the term:
1) statistical literacy - understanding statistics, and even simply understanding that genAI's output is based on statistics, is critical to understanding what AI tools are and how they operate
2) related to this, bullshit literacy - it's absolutely essential to understand that genAI tools are always, only bullshitting, in the technical sense defined by Harry Frankfurt. (It doesn't mean that they don't produce useful output, at times. Just that even the useful output is bullshit, and requires humans to give it meaning.)
Understanding how genAI tools always only parrot, never think or communicate, is essential to critical AI literacy, imo.
Definitely helpful, Matt. We can't understand how LLMs work if we don't have at least a basic understanding of probability. Without that understanding, it can end up looking like search, or creativity, or actually knowledge.
And yes, bullshit, too. I'll need to reread Frankfurt. It's been a while!
"Understanding how genAI tools always only parrot, never think or communicate, is essential to critical AI literacy" -- yes, and a tall order given how prone we humans are to anthropomorphize chatbots
Great post, Chris. I can think of a couple additional elements of critical AI literacy that would be helpful within your definition of the term:
1) statistical literacy - understanding statistics, and even simply understanding that genAI's output is based on statistics, is critical to understanding what AI tools are and how they operate
2) related to this, bullshit literacy - it's absolutely essential to understand that genAI tools are always, only bullshitting, in the technical sense defined by Harry Frankfurt. (It doesn't mean that they don't produce useful output, at times. Just that even the useful output is bullshit, and requires humans to give it meaning.)
Understanding how genAI tools always only parrot, never think or communicate, is essential to critical AI literacy, imo.
Looking forward to future posts!
Definitely helpful, Matt. We can't understand how LLMs work if we don't have at least a basic understanding of probability. Without that understanding, it can end up looking like search, or creativity, or actually knowledge.
And yes, bullshit, too. I'll need to reread Frankfurt. It's been a while!
"Understanding how genAI tools always only parrot, never think or communicate, is essential to critical AI literacy" -- yes, and a tall order given how prone we humans are to anthropomorphize chatbots