[Split thread] Spam, bot, and AI posts

is it just me or has CD seen an influx of rather obvious AI posts recently?

10 Likes

Interesting, a spam ad with no link? I’m just confused at this point.

11 Likes

I feel like getting a #CustomTShirt for some reason. If only I knew where to find one…

3 Likes

I suspect AI is just becoming more and more common in everyday things.

There are a lot of times now where I’m watching a presentation or walk past an ad on the street and can instantly tell it’s AI.

11 Likes

Stock photos in advertisement are being replaced with AI generated stock images which are somehow uncanny valley enough to tell apart from a real stock photo… Which is disturbing that we are that attuned to stock photography now.

I will say we have seen an influx, but I feel like it could be much worse than it is now. I don’t feel like there is much “AI Spam Block” out there yet. Unless the bots are easily identifiable it will get harder to tell with NLP as it improves.

4 Likes

Hot take here: I think all around as a world we hit a peak with generation based AI this week with release of GPT-4o. It’s going downhill after this.

We are hitting the point where AI is starting to learn from itself because AI is so prevalent. As the phrase goes, garbage in garbage out.

Now that Open AI has made their deal with reddit GPT is going to start using GPT generated posts to learn.

I think we will get one more big push when developers are able to teach models how to determine whether or not content was AI generated with high accuracy and then they can clean the database. After that AI will start making content that is undetectable to itself and enter the garbage cycle once more.

Human Drawing of this timeline

All in all: we are safe from the singularity for atleast a couple more months.

13 Likes

I am entirely clueless to recognizing AI, but I haven’t really noticed any uptick in posts that may be AI.

Also, where did this get split from?

split from a deleted post written by AI
you may not notice an uptick because they’re frequently deleted very quickly after being posted

6 Likes

The CD Mod Squad and community flags catch most of it. We (FIRST members) are also pretty wary of new accounts with no team number, and posting stuff that sounds like ads within 15 minutes of making an account.

4 Likes

We are safe until I can see AI open up MS paint and recreate that doodle on its own to prove an argument. That is pure human nature

1 Like

that’s an iphone notes app original

My account was put under automatically before I even posted for being a bot or something like that.

Keep your eyes out. I recently caught one that posted less then once a month (4 posts since December) where the initial posts were innocuous, but were then later edited to add a link that wasn’t obvious to humans (they turned a period into a link).

9 Likes

What could the purpose possibly be though? No one is going to click the link. Is this some strange SEO method?

2 Likes

There was one time my post had to get approved to be seen publicly… what was that? Was it a bot detector?

Serious question: If you were a bot, how would you even know?

1 Like

It could have been a “watched word”. Some of the words in the list are rather innocuous, and we wonder why they were called out. A mod has to either approve it. A mod could also change a word and approve the post. (e.g. change it to “stuff happens” from the more objectionable word)

Do you dream of electric sheep?

12 Likes

Lots of bots are turned away due to CDs robot.txt file that web crawlers are supposed to adhere to. Webcrawlers can ignore this, but it’s like a code of conduct more than actual stopping power. There are other methods beyond this for stopping bots that do not follow the rules specifically called out in the text

https://www.chiefdelphi.com/robots.txt

A robots.txt explanation from CloudFlare can be found here: https://www.cloudflare.com/learning/bots/what-is-robots-txt/

CD Robots.txt Content

See The Web Robots Pages for documentation on how to use the robots.txt file

User-agent: mauibot
Disallow: /

User-agent: semrushbot
Disallow: /

User-agent: ahrefsbot
Disallow: /

User-agent: blexbot
Disallow: /

User-agent: seo spider
Disallow: /

User-agent: *
Disallow: /admin/
Disallow: /auth/
Disallow: /assets/browser-update*.js
Disallow: /email/
Disallow: /session
Disallow: /user-api-key
Disallow: /?api_key
Disallow: /?api_key
Disallow: /badges
Disallow: /u/
Disallow: /my
Disallow: /search
Disallow: /tag/
/l
Disallow: /g
Disallow: /t//.rss
Disallow: /c/*.rss

User-agent: Googlebot
Disallow: /admin/
Disallow: /auth/
Disallow: /assets/browser-update*.js
Disallow: /email/
Disallow: /session
Disallow: /user-api-key
Disallow: /?api_key
Disallow: /*?api_key

Sitemap: https://www.chiefdelphi.com/sitemap.xml

More or less all disallowed

It isn’t just you—there has been a noticeable increase in AI-generated posts across various forums, including CD. This trend is partly due to the advancements and accessibility of generative AI technologies. AI tools have become more sophisticated, making it easier for individuals and even bots to create content that mimics human writing. However, this influx has also led to a rise in the number of posts that are easily identifiable as AI-generated, which can often lack the depth and nuance of human-written content.

This phenomenon is widespread and not limited to forums. AI-generated content has been proliferating across social media and other platforms, leading to concerns about the authenticity and quality of online information. For example, recent reports indicate that even major tech companies are grappling with how to manage the quality and authenticity of AI-generated content, as seen with Google’s recent algorithm updates that penalize low-quality AI content​ (MIT Technology Review)​​ (Search Engine Journal)​.

The visibility of AI-generated posts can often be attributed to the characteristic patterns and lack of personal touch that many of these posts exhibit, making them stand out among genuine human contributions. This shift is part of a broader trend where AI is being increasingly used for content creation, leading to mixed reactions from users and content moderators alike.

2 Likes