Ongoing LLM 'comments bot' malfunction on YouTube | Hacker News Reader

Ongoing LLM 'comments bot' malfunction on YouTube

6 pointsthomassmith654 comments

After a visit to Youtube to check out news on the Israel / Hezbollah conflict, I noticed the comment section is, at the time of this post, 50% malfunctioning LLMs spewing short messages containing random names, mainly English ones.

Presumably the actor's LLM is swapping its fake account names with its message bodies. Does make you wonder how widespread this is; if it's political, the comments are probably confined to middle-eastern conflict videos.

Vaguely interesting. To read them all does make a person entertain the Dead Internet theory.

Examples...

  @SubornaKhatun-b5k
  Brown George Young Jose Wilson Edward

  @BillyDoris-q8k
  Thompson Matthew Hall Angela Young Edward

  @SubornaKhatun-b5k
  Lewis William Gonzalez Kenneth Wilson Jennifer

  @LyttonAtwood-w1y
  Martinez Linda Smith Elizabeth Gonzalez Betty

https://youtube.com/watch?v=5KGuBElCSUU

https://youtube.com/watch?v=lEW7VnHJXd0

Who needs real comments any more? Make them look real and consensus is pushed not formed. This is why human engagement is at a all time low, it's mostly the bots that get the attention with their manufactured support and votes.

And now that TTS is barely distinguishable from an organic person, increasingly more of the videos themselves are 'AI slop'.

So, soon it'll be bots arguing with bots about bot-written, bot-edited, and bot-narrated content.

Yay, slop all the way down!

The ironic part is, it's humans leading the charge, with their heads buried deep in the sand to avoid thinking about how this decays society.

Yes I'm sure we are going to get vast amounts of AI content rubbish everywhere. And for sure we're going to have increasingly serious problems with bad actors using AI in harmful ways.

Others will use AI in positive and creative ways. I think AI-assisted comments can be positive and helpful under correct human oversight, but they are useless if they are surrounded by AI junk.

The AI slop effect is undoubtedly changing the media landscape. And then there will be the adverse AI-related effects that we haven't even got a name for yet, e.g. AI-powered psychological manipulations that subtly influence people, sometimes in personalized individual ways.

Human society will adapt. Perhaps there will be an increased value placed on genuine communications and analog content for example ...

Reply on news.ycombinator.com