It’s getting harder to discern when content is AI-produced. I have to wonder how much current content in the wild is produced by GPT-3 enabled tools.
The purpose of this blog is for me to share the more interesting things I run across, and I will make it clear when content is AI-produced.
I intentionally did NOT do that with the very first post, as an experiment. It is almost entirely GPT-3 written.
I wanted to see how effectively it could masquerade as a human. Not a Turing Test, but in the spirit of it.
If you’re curious to see if you can detect it, stop reading this post now, and browse to https://ghostinthe.garden/2022/09/25/hello-world, “Disruption is Coming” and read through it.
When I read it, I can recognize tone differences between the human and AI content, so it seems obvious to me which is which. I suspect that wouldn’t be the case if I trained a model with my writing style.
In the “Disruption is Coming” post, the title was mine, and the two paragraphs starting with “The conventional wisdom” and “The writing is on the wall” are mine, the rest is GPT-3.
The blog images are all from Midjourney. Going forward any non-human text content will be clearly identified as such in the interest of transparency.
What do you think about the idea of AI-created content taking over the internet? I don’t much like the idea, personally. OpenAI CEO Sam Altman believes AI will produce innovative content and create new ideas. If I start seeing examples of that I may change my tune.