A Human Probably Wrote This
Countless articles on the world wide web are published every day. Many of these are written by us. Some are translated into other languages. Many are unedited, and sloppy or strange. Many are eloquent prose that make the reader consider new ideas from new perspectives. A lot of these are just funny, or mildly interesting. What do all of these articles have in common? Well, people wrote them, of course. Right?
No, of course not. You are an active member of the Internet’s community, and you’ve been around the block. You know computers can write words. But can you tell when an article is written by an AI? Right now, I’m sure you’re skeptical. Maybe an AI wrote this very sentence — would you put money on it?
Are ideas innately human? Can an AI fool you into thinking that an article was written by a human? Or is it more like, well, we all know humans write very, very stupid things, and that there is a fair amount of evidence that humans are better at it than machines. It’s reasonable to doubt that even an intelligent AI, who happens to have good grammar and spelling, would be writing this very sentence. Yes, your computer does write very stupid things, and it is likely that the first thing that you notice is a large quantity of nonsense. But maybe you can’t tell the difference between this sentence and a reasonable human sentence.
There are many methods of writing, and it is hard to gauge whether an AI has chosen the right one. There is no real standard for the quality of writing. A human writer is judged by his ability to make readers think, and the writer who is most successful, in this sense, is the one whose readers think the most. Many readers are concerned with whether or not a writer is trustworthy, and the same is true of the AI. As someone who has spent his entire life learning, my question is: what kind of intelligence to be capable of “the art “ of writing — the kind of intelligence that is most valued in this community?
Not all writing is good, obviously. In addition to being eloquent and thoughtful, good writing is also clear and effective. It is not difficult to write a paragraph of complete nonsense, but how does one write a paragraph that is clear and concise? Without intention, can an AI get a point across? How many questions can I ask in a row?
We’ve all been fooled by AI. But, there is one thing that Artificial Intelligence cannot yet accomplish: fact-checking. This is what got me thinking. You see, I know a lot about animals. I know more about animals than AI. If I wrote out a bunch of outrageous animal facts and fed them to an AI, what would I get in return? Could the responses produced fool other people? What a fun little test!
Introducing: Two Truths and an AI (an animal trivia game.) This is my first JavaScript project. It’s a simple game, there are only six questions. Simply click the response that you believe an AI wrote. If you are correct, you get five points. Sounds like an easy game.
Go ahead and try it out. Challenge your friends! Let me know what you think of the concept, as it’s something I want to expand on in the future. I would love to hear your feedback and ideas for more games like this!
By the way, if you want to know how much of this article was written by an AI, checkout my other post, “An AI Probably Wrote This” here.