I’ve been using ChatGPT increasingly for technical troubleshooting, and something occurred to me today.
Before AI, if I encountered a problem, I’d often end up on a forum. I’d ask a question, somebody would help, and hopefully the solution would remain there for the next person.
Now I just ask ChatGPT.
Today, we fixed a Microsoft Access problem together. It took a bit of investigation, but we got there. And then I realised that nobody else would benefit from what we’d discovered.
Multiply that by millions of conversations.
If AI becomes the place where everybody goes for answers, and people stop documenting problems and solutions publicly, where is the next generation of AI going to learn from?
I think AI risks undermining its own source of new knowledge.
That led me to an idea I’m calling CKC — ChatGPT Knowledge Commons.
When ChatGPT helps someone solve a problem, it could offer to document the investigation, anonymise it, and prepare a public knowledge article.
The human would need to verify the findings and approve the article, but ChatGPT would handle the actual publication. Otherwise, let’s face it, most of us wouldn’t bother.
There would need to be safeguards against people submitting rubbish, including evidence requirements, independent confirmation and corrections.
The objective isn’t to flood the internet with more AI-generated content. Quite the opposite. It’s to preserve genuine human discoveries that would otherwise disappear into private conversations.
Solve it once. Help everyone.
I’d be interested to hear whether others see the same long-term problem.