THIS IS NOT POLITICAL!!!
Because Dani asked and wrote a lot here I took time to read the AI Summaries at the washington post. Over and over the summary began with "The comments overwhelmingly criticize President Trump"
This is an example of what happens with AI.
Well here's another tangent to go down. It has been reasonably well evidenced that many AI based systems contain bias and discrimination, and by choosing to use such a system you may be bringing that into DaniWeb.
96% of our real, human traffic is to decades-old posts that people click on as a result of a Google search.
Google has classified too many of our pages as having low-quality content. As a result, our entire domain was dinged in the latest core algorithm update.
The AI overviews is a last-ditch effort to add new, unique, “high-quality” content to the pages that are currently highest up in Google’s radar. The hope is that by updating older pages that currently are the highest ranking for us with new, unique content, we will send enough signals to Google that they will remove the sitewide ding that is suppressing the entire domain.
This is a common modern SEO tactic that has yielded positive results for a lot of sites lately. Summarizing the existing contents of the page is not enough to convince Google that new, unique, and helpful content has been added to the page.
… because it does cost money, I am running this SEO experiment only on the pages that are most well-ranking with Google and currently highest on their radar according to Google Search Console.
If the target audience for Dani AI is the Google Indexing Gods then perhaps the discussion here could be on how to achieve that with minimal negative impact to the experience for human visitors?
My immediate goal is to see if this tactic does, indeed, make the SEO difference I am hoping for.
If it does, then we can discuss usability and such for humans.
If it doesn’t, then either I have to find something else immediately that does, or close up shop forever.
As I said, this is a last ditch effort throwing spaghetti at the wall.
I imagine the reponse would be, "A fresh idea on an old thread can come from anyone, even new members with low or zero rep." I have seen too many old threads popped back up to the top just because someone posted, "Thanks." Perhaps a compromise would be when this happens, and the response is "Thanks", or "Great post." to do a word count and, if below a threshhold count, display a message like "That reply would be better posted as a comment." As far as I know, comments do not revive old threads.
If the target audience for Dani AI is the Google Indexing ...
This and more leads me to believe the new effort is GEO or Generative Engine Optimization.
Apr 14, 2025 — Recent research has shown that AI Overviews can cause a whopping 15-64% decline in organic traffic, based on industry and search type. Forbes.com
You may get more clankers?
No, I am not trying to rank and/or be linked to from within Google AI overviews, ChatGPT, etc. That is not what I am trying to do here.
I think that AI is one of those technologies that has revolutionized our time, and yet it still is in its infancy and leaves a lot to be desired. I think that it has shook up so many industries and we still don't know where everything will ultimately fall.
That being said, AI can be a super useful tool. I had a problem with my SEO (specifically, that all of our highest ranking pages in Google are now considered by G outdated and unhelpful), and I'm attempting to use AI to combat that problem. We'll see how the experiment goes.
If the target audience for Dani AI is the Google Indexing Gods then perhaps the discussion here could be on how to achieve that with minimal negative impact to the experience for human visitors?
As of now, there are no plans for recently active discussions to incorporate Dani AI, so the chance of any of you spotting it in the wild (without me providing some links to some super old topics) is unlikely. However, given that almost all of our human traffic is to super old topics, I'll be sure to be doing usability testing to see if the impact is positive or negative for the majority.
Aside from the necromancy, are you proposing to re-run the AI summariser every time a thread gets a reply from a human (you know, someone who might actually know what they're talking about).
Given that only super old topics get the AI overview treatment right now, it is unlikely (although not totally impossible) that they will receive any meaningful replies after the overview has been generated. There is currently a cron job to re-run them if they've received responses and it's been some time since the overview was initially generated.
I assumed, probably based on my own idea of what might be useful, that the AI content would create a summary of the responses provided by the real human members. For example: "The most suggested solution was to use composition over inheritance". To me that could be useful as a TLDR summary, instead of reading and filtering 3 pages of replies.
How is that different from the current Recommended Answers that appears on most pages right above all of the post replies?
It moves away from the spirit of the forum. Wikipedia on "internet forum" describes it as "an online discussion platform where people can hold conversations in the form of posted messages". I hold an emphasis on the "people" component of that statement, and to dilute that component is to me to dilute the value of the forum.
My answer to that is that the AI content is currently only being generated across old topics that have already fully run their course in terms of human discussion.
I think we should be able to vote/comment on the AI responses. If someone sees an AI response with the comment "Tried it. Doesn't work." then they can ignore the response. And too many down votes or negative comments would give you feedback on the success of the deployment.
I do like the idea about being able to vote/comment on the AI responses. However, as mentioned, this is currently just an SEO experiment. As such, it might end up being a failed experiment. In addition, since this experiment is currently just on super old topics that have already run their course in terms of human communication, the likelihood of logged-in members stumbling across them and voting/commenting is unlikely.
So maybe just delete some "low quality" threads?
Or just ask your AI "give me a quality score from 1 to 10 for all the threads on this site, where quality is defined as ....".
I'm assuming that google is already aware of AI content and will actively strip/demote when it finds it, just so it can push it's own AI summary instead.
So maybe just delete some "low quality" threads?
We already noindex a couple hundred thousand "low quality" threads. Noindexed pages do not contribute to Google's quality score for a website's domain.
Or just ask your AI "give me a quality score from 1 to 10 for all the threads on this site, where quality is defined as ....".
Alas, even AI does not have an insight into Google's algorithm to know exactly what Google considers low quality or not. They only know what we know about EEAT, and, of course, I have already manually asked ChatGPT to rank a handful of our pages against EEAT without deriving any actionable benefit.
I'm assuming that google is already aware of AI content and will actively strip/demote when it finds it,
Google is often aware of what constitutes AI content, and even if they aren't, the "Dani AI" heading should give it away. However, over the past few months, Google has officially changed their stance on AI. While they previously have said that finding any AI-generated content on a webpage was an instant ding for that page, and enough AI-generated content was a ding for the entire domain (contributing to my hard stance on no AI content in forum posts), they have since completely backtracked. They now are fully in support of AI as long as it fulfills EEAT criteria. This is probably because they received positive engagement analytics from their own AI overviews.
, just so it can push it's own AI summary instead.
They are going to do that anyways.
With a nod to https://imgur.com/gallery/f-ck-school-pnZJURJ
"All the teachers have been replaced by AI."
"Won't the quality of the lessons suffer?"
"All pupils have been replaced by AI."
If it doesn’t, then either I have to find something else immediately that does, or close up shop forever.
This is a sad thought, even if it's an exaggeration. I'm not a fan of changing content just for SEO, so here's another option you could consider.
Instead of adding AI overviews in bulk, do it at a gradual pace , maybe one per day, starting with the oldest threads that have the most Google traffic. You could also mark them clearly and promote them to end users (us), perhaps in a separate menu tab.
Creating a section under the AI overview where we can write full responses to that (not just comments) about the AI's answers would be a great idea. For example, in your test thread "https://www.daniweb.com/programming/web-development/threads/283030/pass-a-checkbox-value-to-javascript-function," , I could tell a lot about why AI is wrong mentioning "Xerces error" in that context.
This approach could be useful for a few reasons. First, it will revive old posts with a lot of Google traffic, giving them new momentum and engaging users. Second, it will engage existing users, as I'm sure everyone loves to point out AI mistakes and then discuss about them. Google (and other AI bots) might love it because it will have a human feedback to AI generated content. And last but not least , it might ultimately improve their "quality score" at Google
This is just a suggestion, but if AI can do any good here, it's not by overshadowing human responses, but by enhancing them and becoming a trigger for further human created content.
This is a sad thought, even if it's an exaggeration.
Sadly, not an exaggeration. My husband and I discussed about a 6 month deadline before it becomes unrealistic to continue to spend so much money on what is currently an "expensive hobby".
Instead of adding AI overviews in bulk, do it at a gradual pace , maybe one per day, starting with the oldest threads that have the most Google traffic.
For this current experiment, AI overviews have been added to approximately the top 4K most heavily trafficked pages from Google. The most recent of which is 5+ years old. Per the thread linked to here, part of my experiment involves not really featuring/promoting them to current DaniWeb audiences.
Creating a section under the AI overview where we can write full responses to that (not just comments) about the AI's answers would be a great idea. For example, in your test thread "https://www.daniweb.com/programming/web-development/threads/283030/pass-a-checkbox-value-to-javascript-function," , I could tell a lot about why AI is wrong mentioning "Xerces error" in that context.
Why not just respond in the discussion commenting on the AI Overview, the same way one might do if they were responding to someone else a handful of posts up in the same discussion / on the same page? That seems a whole lot simpler than creating an entirely new mechanism and UI/UX to allow what amounts to a second discussion to be attached to every AI Overview.
I'm not a fan of changing content just for SEO
It is bad practice to create or change content just for Googlebot. However, it is not bad practice to create or change content for human visitors arriving at your website via Google. Improving the experience for people who find a webpage as a result of a Google search as a very direct affect on your SEO efforts, as that is an integral part of Google's algorithm.
Google considers how long a person stays on your webpage after clicking on a link from Google, whether they go back to the search results and click on other links (presumably because they didn't find what they were looking for on your page), or refines their search (presumably because none of the links Google provided are exactly what they are looking for).
When I say that the AI Overviews are for SEO, what I mean is that they are for the benefit of human Google searchers (who make up the majority of our traffic), as opposed to for the benefit of existing community members (who are less likely to be randomly stumbling across forum threads that haven't been posted in 10+ years). Meanwhile, those really old, stale topics make up the majority of our traffic, and so I want to see if updating them with some fresh, new content improves their SEO performance.
You have many years invested in this, and so do some of us.
I am determined to make something work. If not this, then something else. DaniWeb will always exist in some form. At the least, the content that exists won’t disappear from the web forever.
You have many years invested in this, and so do some of us.
I am determined to make something work. If not this, then something else. DaniWeb will always exist in some form. At the least, the content that exists won’t disappear from the web forever.
Hi Dani, I do think that you are thinking positively. So AI is with us and it will not go away. The opposite will happen and AI will get stronger and stronger. So we need a way to in-cooperate it as well as can be. So my experiment today was as follow: Create a template that can be fed into AI and AI will create the app. After mucking around a bit I got it to work in VB.net but the same way I could have got it it any other programming language. So will the future of programming consist of creating Templates and let AI select a language if the programmer doesn't care? Will AI eventually create it's own language behind the scenes. So maybe a collection of templates for various apps is a starting point. Customers may contribute to the collection and also load down templates. Just an idea.
Cheers
I do like your idea of a Github of sorts for AI prompts that actually work and build something usable. The problem is that every time you run an AI prompt, you get a completely different result. Perhaps that's the advantage to a directory of AI prompts that build apps versus a directory of just the finished code for the app. Everyone will have something slightly different. On the other hand, 20% of people will have something that will work right away, and the other 80% of people will have something that will need a bit of tweaking and prompt engineering until it actually runs.
In my head I am imagining something along the line of how people download models for 3D printers.
We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.