ChatGPT-3: Google Says AI Generated Contents Are Against Webmaster Guidelines
steadipulse.com
steadipulse.com
Google's official position was published on 8 February here:
https://developers.google.com/search/blog/2023/02/google-sea...
It's a much more nuanced position that can be summarized as "make sure you create good content, however you create it". A focus on quality, not process, is reasonable.
Disclosure: ex-Googler in search.
About the news article title, looks like it's trendy to make Google seemingly worried about chatGPT. and it's been a trend to claim google penalises its potential competitor via the means it can (search).
With google quality content challenges, it makes sense they would focus on quality, and if AI generated is what make search users click more that's what it will feed most, especially if that's promoted content, as it has always been.
If Google, Facebook, CNN, or Buzzfeed wants to do it I’m sure it’s completely approved and Google will even encourage it by ranking them higher in the search results.
They just don’t want these tools in the hands of ordinary people, that’s what they always mean when they use words like “Safety”. Safe for them!
She's just been hired by a client to copywrite ChatGPT-generated blog content to make it "more human." Ordinarily it takes her about 3 hours to author a blog post for a client. Now it takes her 30 minutes to "fix up" a ChatGPT-generated blog post, saving the client about 2.5 hours of copywriting expense.
She's beside herself that she's no longer getting paid to author content but is instead fixing up machine-generated content so that Google doesn't penalize the ranking. I get the sense that we are now in a cat-and-mouse game between the AI models used to generate content and the AI models used to detect AI-generated content, and people like my friend will soon be out of a job entirely.
Of course that doesn’t address the underlying emotional impact of the job changing from authoring original content to proof reading for a robot.
Content creation is not a good business to be in now.
ChatGPT, please write a story about a unique word that has never been used before. I want to see if it ends up in Google.
ChatGPT:
Once upon a time, there was a small town called Schmorgleville. The citizens of Schmorgleville were a strange bunch, and they had an even stranger custom. Every Saturday night, they would gather in the town square and dance the schmorgle.
No one knew exactly where the schmorgle originated from, but it was said to be a dance that had been passed down from generation to generation. It was an odd dance, full of twists and turns and hopping up and down. Everyone in the town seemed to love it, though, and would dance until the sun came up.
One day, a stranger came to town. He was an old man with a long white beard and he seemed to have a great knowledge of the schmorgle. He taught the citizens of Schmorgleville all the steps of the schmorgle and soon, they were all performing it with enthusiasm.
The old man stayed in town for a while and enjoyed watching the citizens of Schmorgleville dance the schmorgle. When he finally left, he told them to keep the schmorgle alive and to never forget the joy it brought them.
And so, the people of Schmorgleville still dance the schmorgle every Saturday night. They call it the 'Schmorgle Dance' and it is a reminder of the old man and the joy he brought to their lives.
It’s been on a steady decline for a long time with dark SEO and filler content patterns.
Google already tries to just give you the right answer to your question and not send you to garbage websites.
Someday we’ll seldom need to read any websites directly, especially for anything that content marketing and think pieces are trying to answer.
If you want that, chat with the AI to ask questions and get answers directly. Ask more questions to get more angles of thought.
This is a good thing IMHO. People using AI to fill their websites with content is a waste of everyone’s time.
The big question is where does all the training data come from if not from scraping the whole web.
We will figure out some sort of provenance system for data, which will lend to the high quality data sets of human curated knowledge and art.
These will have trade offs, like limits to the knowledge and more expensive to build and consume. This will compete with data sets that do scrape everything.
What’s ironic is we’re shifting back to the days of Yahoo curating a directory of content, which Google ultimately destroyed with their crawling and ranking.
Everything is cyclical.
Funny, because every goddamn time I try to find a solution to a very specific issue, the first page is full of sites that repost AI-mangled crap that tells you to keep your computer and firewall updated.
Hoating that hot garbage ought to be punishable.
Instead of giving me what I want, it gives me pages of useless content farm crap with titles such as "How to Fix WHATEVER.dll Not Found or Missing Errors", "WHATEVER.dll free download", etc, etc, etc.
I can't understand why Google can't expunge all this content farm junk. I don't believe it is technically infeasible – rather, I think it is that Google (for whatever reason) doesn't actually want to do so.
Not good business to send you directly to the ad free reference site with all the answers.
If i had to bet, Google won’t survive. Someone much smaller and nimble will take over.
Right now we back in the “mainframe” days where IBM, MS, and Google have the systems.
Next will be “super computers”, then “work stations”, then “personal computers” then “mobile”. At the end we’ll all have many data sets and AIs in our pocket.
This cycle will be way faster because we don’t have to invent the underlying computer tech.
You can already get the AI in your pocket to classify and content search all you photos.
More like a shift back to universities producing, curating and verifying the highest quality human produced data.
But this is interesting: If google excludes AI content, then they are opening themselves to be sued for monopolistic behavior by their competitors, no?
Google search is full of content farm crap, and I don't believe Googlers aren't smart enough to work out how to stop it, if they were truly motivated to do that.
Maybe, someone has decided that content farm crap is better for advertising revenue than real content. Or that the "average user" would rather crap than real. Or that the expected gain in advertising revenue from improving the search is less than the cost of doing so.
The ad platform was always manually reviewed and was never allowed to have automated writing. Declaring AI written content as automated is hardly news.
Lets say in a couple of years the web and any other global content output gets to a share of 10-20% of “AI” generated content, that would definitely change what it spits out. The question is would that be bad or good?
So in a sense it could be we are now at peak ChatGPT output quality, since the inputs given to it are 99% from humans.
One interesting thing is that the situation you describe provides a huge moat for the first successful system. For instance, OpenAI can store and fingerprint all of the output from GPT-3 and ChatGPT. They can use these fingerprints to prevent training newer versions of GPT on old outputs. Less popular systems won't be able to sanitize the training data as well.
Certainly we already need tools for this just like those to detect plagiarism.
Cat and mouse game I know but AIs that can detect AI content seems like a reasonable outcome to me.
https://developers.google.com/search/blog/2023/02/google-sea...
Unless "Steadipulse.com" is now the new canonical source for Google's webmaster guidelines, AI content is fine as long as it's human reviewed. Spam is still spam, and they consider unreviewed/unedited AI content to be untrustworthy.
Google Cloud's translation API also has a TOS that requires you to tag the output as machine translated if you're publishing it unedited on the web.