Comment spammers use AI in another assault on bloggers

2 October 2025

When I turned comments back on here a few months ago, after an absence of many years, I was amazed at how quickly spam comments began appearing. Good news travels fast it seems. A new outlet has appeared for us to post our drivel — quick — get over there. But because every comment made here is held back for approval, none of them ever see the light of day.

Of course I wasn’t really surprised at the speed at which the spam arrived. Nor the lack of genuine comments, though there have been a few. I re-enabled comments as a way to centralise my web presence back onto this website. I’m not the biggest fan of social media, centralised or decentralised, but not because I dislike it (well, not too much), rather social media is just too time consuming.

What did dumbfound me though was the empty-headed nature of the spam comments being left. Some were barely coherent, while others were literally single words made up of random letters. What blogger, in their right mind, is going to approve those sorts of comments? A time-poor blogger, or one not paying attention, I think might be the answer.

These senseless utterances aren’t offensive, so maybe they’ll, you know, just get approved. And with some websites allowing follow-up comments from the same person to be posted without moderation, the floodgates would be open. But I suspect few spam-commenters saw much of what they wrote ever approved. But now they have changed tactics, and are using AI to craft their foul fare.

A lot of the recent comment spam I’m seeing looks as if the writer has read the post they’re responding to, through the way a comment is worded. When I posted about Tim Berners-Lee a few days ago, a lot of comments similar to this began appearing:

Oh Timothy, your call to have AI development moved under the auspices of a global not-for-profit is just a little simplistic, don’t you think? Yes, you invented the internet and gave it to us for free, and for that I thank you from the bottom of my heart. But placing AI development in the hands of a non-commercial entity is asking an awful lot.redacted spammy link

At first pass, the comment seems genuine. I too thought Berners-Lee was being optimistic in the extreme by suggesting a global not-for-profit organisation oversee future development of AI, but I’d never call Berners-Lee naive. He knows what he’s saying, and the idea makes sense, though I can’t see it ever happening. But that’s another story.

A commenter though is entitled to their opinion. And it almost seemed like an actual point-of-view, but for the ridiculous inclusion of an embedded spam link. Without, notably, a space after the previous sentence. The writer seems switched-on, but their oddly deficient syntax betrays them. And then the question: why on earth embed a spam link within the comment?

Did they not see the field on the comment form that allows a URL to be included? It’s possible I might have missed the spam-link if they did that. Usually though, I look closely at the URL of a commenter’s website. But then going on to post numerous, slightly differently worded, variations of the same comment, from the same IP no less, somewhat gives the show away.

Even writing this article is helping train the AI spam-commenters though. What bloggers, who allow comments, are facing though are somewhat more sophisticated spammers, who are using AI to compose comment spam that look like the real deal.

And yes, I look forward to seeing the thoughts of the AI spam-commenters in response to this post.

RELATED CONTENT

artificial intelligence, blogs, technology, trends