If an ultra-sophisticated language algorithm can pass a college writing assignment, that's a sign of a problem, but it's not necessarily a problem with the AI, or the assignment. h/t to @BryanAlexander for the tip on the original article. insidehighered.com/blogs/just-vis…
When @BryanAlexander sent me the original article, I thought I would have a very different take, that I would be either skeptical of the AI or critical of the nature of the assignments. Neither turned out to be true.
The AI (GPT-3) is pretty freaking amazing, like how Deep Blue was amazing when it beat Kasparov. Language AI couldn't do what GPT-3 does until GPT-3 did it. The seemingly impossible is suddenly possible.
But it's important to note that the algorithm did not ace the college assignments. It passed them. The reason it passed them is the thing we need to look at more closely and it's wrapped up with how we assess writing.
For more than that, I ask you to read the post, but as a mild spoiler #ungrading. insidehighered.com/blogs/just-vis…
• • •
Missing some Tweet in this thread? You can try to
force a refresh
