Reddit using Large Language Models (LLMs) to ban Accounts Now

Reddit says it developed tools with LLMs to cut down on spam, much of which was created with LLMs in the first place. It’s a bit ironic, but in the AI era, platforms have no choice but to fight fire with fire. According to the platform, Reddit blocks 23 million spam views per day and catches about 25,000 new spam posts and comments each day.

Social platforms have been building automated spam reduction tools for years, but Reddit says these updated tools are catching spam at a higher rate.

“We leverage LLMs to catch the highly subtle, coordinated patterns of fake behavior and artificial hype that older systems once missed,” a Reddit blog post says. The company claims it reduced users’ exposure to spam by 20% from January to March compared with the prior three months.

Click to expand...

exceprt taken from:
https://techcrunch.com/2026/07/06/reddit-is-using-llms-to-solve-a-problem-llms-largely-created/



And this includes noticing real human patterns vs AI content patterns.

Within the LLM models is tech to noticing AI-generated content and ban it.


So what's a layman's summary of what a Large Language model is?

Essentially it's trainable Ai models that learn and adjust to human writing habits and filter out spam, toxicity, inccorect spelling and grammar, marketing language, etc. These tools can compare patterns to understand what's 'generally' natural content , what's apppreciated and what's not. Then it can also research Ai responses and compare it to natural response.

Once the model can classify certain types of posts, they can classify it in a particular category.

And those categories will determine if the post is considered fake or spam, or other. And from there, posts get flagged . removed, or all-together blocked.

It's been known for quite a while

bartosimpsonio said:

It's been known for quite a while

Click to expand...

I shoudln't have put 'now' in title maybe but moreso there's changes that are being noticed with the LLMs now to successfully net a wider audience of negative content and Far More Effective

Regardless, this is headline news this week. Where did you see it before?

game getting harder then before

DojaCat said:

game getting harder then before

Click to expand...

Exactly.

For those using automated tools or that do not understand english fluently (Reddit's majority audience by far),
bieving this is 'easy' news,
does not understand what's going on.

This is vital news.

23 million spam views? Those are crazy numbers. It seems we're not the only ones interested in marketing or growing on reddit.

HenryObi said:

23 million spam views? Those are crazy numbers. It seems we're not the only ones interested in marketing or growing on reddit.

Click to expand...

They blame us for nothing. The real spammers wear white hats and blend in with the crowd. The ones who survive for long become digital marketing gurus.

I think this is a positive move. The days of low quality spam and artificial hype are coming to an end. Brands that focus on genuine discussion and adding value to communities will have a much better chance of succeeding on reddit

TheStrategos said:

I think this is a positive move. The days of low quality spam and artificial hype are coming to an end. Brands that focus on genuine discussion and adding value to communities will have a much better chance of succeeding on reddit

Click to expand...

genuine discussion and adding value? do you hear yourself? these are brands, with marketing in mind. these things do not exist here.