Dead Internet Theory because AI search results are good enough

[David777]

The below is rather funny, at least to this person. The rate of technology change is devouring some of its own. That is web advertising click reduction due to people being content from Google AI and other web search AI summaries without having to actually drill down and select link hits for websites that Google AI got their information from freely and is now using as AI summaries that reduces probability searchers will bother to visit those websites. Websites that generate ad revenues from numbers of Internet searches. Now some are suing Google.

Is the Dead Internet Theory Coming True?

Of course, there are myriad web sites that act as information sources that tend to grow and grow as more people use them and then over time they expand whatever information. The Internet has had continual and unpredictable technology changes that have made it evolve. Many of the changes have been the result of increasing bandwidths that allow more capabilities. The biggest change so far has been the rise of smartphones and the Mobile Internet. Also for instance, the change from black and white images to tiny color images then gradually larger color images. Or the rise of YouTube when video codecs and bandwidths increased enough for video. Much more, of course.

 

[quintinmorrow]

Lately it really does feel like fewer real people are on the open web, or at least fewer clicks are happening. AI summaries answering everything upfront changes habits fast, especially for casual browsing. That doesn’t mean the internet is dead, but it’s definitely shifting who creates content and who actually gets seen anymore.

 

[MoBeans]

I did a google AI summary on myself and there wasn't a true statement in it. It had my mother in-law as my wife and my wife as my daughter. It said I was in the solid waste management business and was a former head chef in a local restaurant. I shutter to think someone out there is making important decisions using it, like whether to hire a person or not. Now google is putting ads in their AI so if you ask something like whats the best shampoo you get a list of sponsored products.

More than 51% of all internet traffic is now automated (as of 2025 reports from Imperva and Cloudflare). AI bots are now creating content (posts, images, articles), other bots are "liking" and commenting on them to boost engagement and manipulate opinions. And search engines are indexing that bot-created content to train new AI. Aside from poisoning the AI training data this creates a closed loop where humans are targets instead of creators.

Even when you see a human post, it was probably chosen for you by an algorithm designed to maximize profit, not to foster genuine discussion. A separate study by Ahrefs in early 2025 looked at 900,000 newly created webpages and found that 74.2% of them contained some form of detectable AI-slop You can't trust anything anymore on the internet. I created my first govt. website in 1992 (actually it started as a gopher server) and worked in the technology field for 25 years. The internet I grew up with, populated with useful, trustworthy information is already dead.

 

[Medusa]

I did a google AI summary on myself and there wasn't a true statement in it. It had my mother in-law as my wife and my wife as my daughter. It said I was in the solid waste management business and was a former head chef in a local restaurant. I shutter to think someone out there is making important decisions using it, like whether to hire a person or not. Now google is putting ads in their AI so if you ask something like whats the best shampoo you get a list of sponsored products.

More than 51% of all internet traffic is now automated (as of 2025 reports from Imperva and Cloudflare). AI bots are now creating content (posts, images, articles), other bots are "liking" and commenting on them to boost engagement and manipulate opinions. And search engines are indexing that bot-created content to train new AI. Aside from poisoning the AI training data this creates a closed loop where humans are targets instead of creators.

Even when you see a human post, it was probably chosen for you by an algorithm designed to maximize profit, not to foster genuine discussion. A separate study by Ahrefs in early 2025 looked at 900,000 newly created webpages and found that 74.2% of them contained some form of detectable AI-slop You can't trust anything anymore on the internet. I created my first govt. website in 1992 (actually it started as a gopher server) and worked in the technology field for 25 years. The internet I grew up with, populated with useful, trustworthy information is already dead.

I have caught AI giving me incorrect answers many times now.

My son told me to, when I let AI do the thinking for me, imagine the wrinkles in my brain matter being ironed into smoothness.

 

[Paco Dennis]

I visit "BlueSky" social gathering site that was started to combat the commercial take over of the internet. Advertising itself on authentic content, and putting down the main stream media.

I just checked to see how much that is true. It is loaded with bots of every kind read this report.

It is difficult to give a single "official" percentage for Bluesky specifically because the platform's decentralized nature makes it harder to track than a traditional site like Facebook or X (Twitter).

However, we can look at the current 2025/2026 data to get a very clear picture of the situation:

1. The "51% Rule" Still Applies​

While the 51% bot traffic statistic you cited (from the 2025 Imperva Report) covers the entire internet, social platforms are often the most concentrated targets. On Bluesky, while the platform is still smaller than X, users and researchers have reported a significant surge in bot activity as the platform crossed 40 million users in late 2025.1

2. High "Impersonation" and Engagement Rates​

A study conducted in late 2024 and early 2025 found some startling trends on Bluesky specifically:

• Duplicate Accounts: Roughly 44% of the top 100 most followed accounts on Bluesky had at least one bot-driven duplicate or "copycat" account posing as them.2
  
  
• Engagement Spikes: During periods of high user growth (the "migration waves"), automated sign-ups have peaked at rates of 6 to 10 users per second, many of which are identified by community moderation as bot accounts designed to "juice" stats or follow-spam real users.

3. The "Anti-AI" Culture vs. The Reality​

One unique thing about Bluesky is its community's strong anti-AI sentiment.

• Community Shields: Because Bluesky allows "Labelers" (community-run moderation tools), many users subscribe to blocklists specifically designed to hide "AI Slop" or accounts with AI-generated profile pictures.3
  
  
• The Infiltration: Despite these efforts, reports in mid-to-late 2025 indicate that "AI-girls" (bot accounts using AI-generated photos to lure clicks) and political "reply-guy" bots have become a major annoyance, appearing in as much as 20-30% of comments on high-traffic political posts.

---

Summary Table: Bot Presence on Bluesky (Est. 2025-2026)​

Type of Activity	Estimated Percentage	Description
Global Web Traffic	51%	The total internet baseline (Imperva/Cloudflare).
Top Account Duplicates	~44%	Bots posing as famous figures to siphon followers.
Comment/Reply Noise	15–25%	Varies by topic; highest in politics and crypto.
Daily Active Users (DAU)	~10–12%	Only about 4M of the 40M registered users are daily actives; many "ghost" accounts are likely bots.

The bottom line: While Bluesky feels "cleaner" than other platforms because of its blocklists, the sheer volume of automated traffic trying to get in is consistent with the rest of the web. It is less of a "dead" platform and more of a "besieged" one.

 

[retiredtraveler]

"I have caught AI giving me incorrect answers many times now.".
Yeah. I like using AI for detailed info on hiking routes for travel overseas along with other travel info. But I have to double check the info against 'regular' websites to be sure the info is correct. And the info is not 100% in some cases.

 

[outlander]

The biggest issue of an internet brimming with incorrect information? Is that most people today believe it, since they can't think for themselves