Vice recently published an article about ChatGPT generated spam on the popular AskHistorians subreddit. Here’s an excerpt:
"The two-million-strong AskHistorians forum allows non-expert Redditors to submit questions about history topics, and receive in-depth answers from historians. Recent popular posts have probed the hive mind on whether the stress of being “on time” is a modern concept; what a medieval scribe would’ve done if the monastery cat left an inky paw print on their vellum; and how Genghis Khan got fiber in his diet.
Shortly after ChatGPT launched, the forum was experiencing five to 10 ChatGPT posts per day, says Gilbert, which soon ramped up as more people found out about the tool. The frequency has tapered off now, which the team believes may be a consequence of how rigorously they’ve dealt with AI-produced content: even if the posts aren’t being deleted for being written by ChatGPT, they tend to violate the sub’s standards for quality."
- The reason AskHistorians was targeted is the simple Q&A structure – this makes it very easy to input a question as a prompt to ChatGPT and get a plausible – and possibly factual (but who knows!) answer out.
- Spam has always been with us, and the race to moderate spam has been a long running one. I think we’ll have AI systems that moderate spam that will catch 80% of it, but the other 20% will be smarter than the systems and need to be caught by humans. That work will then inform the next generation of AI systems that catch spam. And so on, ad infinitum.
- I hate to call this the demise of social media, but if spam gets out of hand it will be. We’re already seeing a lot of online conversations move to newsletters and private Slack communities. I think ChatGPT generated spam will cause that trend to continue. I personally welcome the newsletter approach, because it harkens back to the “good old days” of the web where people built relationships and connections with one another based on blogs. Maybe blogs will make a comeback?
What do you think?