How does Google identify AI content?

[paisawaisa]

Some members on BHW have mentioned that AI content might have got punished by Google. So my question is how does Google identify AI content?

 

[Richichi]

There have been countless posts about that.
AI uses same phrases again and again and uses a very easy-to-spot wording in the sentences, really not hard to detect.
Google said they might punish AI content that gives false information, this could be checked with a comparison to Wikipedia or something

 

[howard981]

Just google openai detector, it's the first result.

You can see how easy it is to spot that any ChatGPT generated content is over 99% fake/AI generated.

Now, that's an open-source solution. Now imagine what a trillion dollar company can do.

 

[CodingDisorder]

Google doesn't give a shit, there are TRILLIONS of websites on the web, why should they waste GPU power to go after AI?
The most important stuff is for your articles to provide value. That's it.

 

[howard981]

Google doesn't give a shit, there are TRILLIONS of websites on the web, why should they waste GPU power to go after AI?
The most important stuff is for your articles to provide value. That's it.

That's just your hunch. It's quite the opposite in reality.

Here's what Mueller, a high-rank Google webmaster says:

For us these would, essentially, still fall into the category of automatically generated content which is something we’ve had in the Webmaster Guidelines since almost the beginning.
And people have been automatically generating content in lots of different ways. And for us, if you’re using machine learning tools to generate your content, it’s essentially the same as if you’re just shuffling words around, or looking up synonyms, or doing the translation tricks that people used to do. Those kind of things.
My suspicion is maybe the quality of content is a little bit better than the really old school tools, but for us it’s still automatically generated content, and that means for us it’s still against the Webmaster Guidelines. So we would consider that to be spam.

Search for: "how will google treat ai content" to learn more.

 

[spectrejoe]

Google doesn't detect AI content. Google detects low-quality content. There's a big difference.

 

[howard981]

Google doesn't detect AI content. Google detects low-quality content. There's a big difference.

No one knows how exactly the proprietary software works or what it does [except those with access to the source-code].

You think Google will not notice a proliferation in blog articles in the last few month?

And not realize its driven by OpenAI solutions?

Do you think their algorithm never changes?

Do you think there isn't a decent chance they might penalize/categorize content as AI generated?

 

[HyperLesso]

Some members on BHW have mentioned that AI content might have got punished by Google. So my question is how does Google identify AI content?

Probably of the algorithm ai use for every article, ye I don't know what it is; just my thought

 

[spectrejoe]

No one knows how exactly the proprietary software works or what it does [except those with access to the source-code].

You think Google will not notice a proliferation in blog articles in the last few month?

And not realize its driven by OpenAI solutions?

Do you think their algorithm never changes?

Do you think there isn't a decent chance they might penalize/categorize content as AI generated?

Yes

Yes, although OpenAI is not the only text AI that exists.

Yes, but I can say with 99.9% certainty that they will never truly detect AI content, even public ones like OpenAI which leave big footprints. AI Detection is very expensive and time consuming when done at scale, Google goes through BILLIONS of articles every day, and they can't afford to spend extra time & resources to detect if it's made by AI or not.
- Costs them extra money
- They have been reducing crawl rates over the past 2-3 years because of huge costs already, they aren't gonna exponentially increase their costs
- AI detectors are not as good as you think. Plenty of legit human content also gets flagged by AI detectors, the implications of penalizing so many innocent people IS MASSIVE.
- Extra time spent detecting AI (which is actually time expensive) means even less crawl rate on their bots which have already seen quite an effectiveness reduction over the past few years and will continue to happen as the content on the internet grows and grows and grows.

 

[howard981]

Yes

Yes, although OpenAI is not the only text AI that exists.

Yes, but I can say with 99.9% certainty that they will never truly detect AI content, even public ones like OpenAI which leave big footprints. AI Detection is very expensive and time consuming when done at scale, Google goes through BILLIONS of articles every day, and they can't afford to spend extra time & resources to detect if it's made by AI or not.
- Costs them extra money
- They have been reducing crawl rates over the past 2-3 years because of huge costs already, they aren't gonna exponentially increase their costs
- AI detectors are not as good as you think. Plenty of legit human content also gets flagged by AI detectors, the implications of penalizing so many innocent people IS MASSIVE.
- Extra time spent detecting AI (which is actually time expensive) means even less crawl rate on their bots which have already seen quite an effectiveness reduction over the past few years and will continue to happen as the content on the internet grows and grows and grows.

I get it, but I don't fully agree.

They will invest in it, if they deem it profitable.

AI generated content detection isn't as consuming as you think.

Take a look at my first post in this thread.

All I am saying, it's not a 0% chance that they will not implement it.

So if you are trying to build purely AI content sites, those people
are taking huge risks right now.

 

[ranga]

The debate is over 2 point and 1 decision:

1. Google can/cannot detect AI content reliably without penalising human content: IMHO they have much more advanced system than we can imagine, so they can do it if necessary keeping error rate low. Error rate will be there but Google has to bring it in acceptable levels. for e.g. spammy link building has an .error rate and websites do get hit with negative seo.

2. Google can detect but don't have resources: IMHO Google is a behemoth and if AI content start threatening their business model, believe me they will find necessary resources.

Now the decision:

1. Will Google Ban AI Content if they have capability: IMHO their content quality strictness will increase, they can't just ban content if its AI written if it adds value. They will demote it if they find content with better depth than AI content. AI content itself is not bad, if proper research is done before writing.

If they start banning, it will become cat and mouse game like SEO, people will tweak their content to avoid AI detection (ultimately improving the quality).

Also, AI is the next disruption (it can make search engine itself irrelevant), so they can either embrace the change or be eliminated like Nokia

 

[paisawaisa]

Just google openai detector, it's the first result.

You can see how easy it is to spot that any ChatGPT generated content is over 99% fake/AI generated.

Now, that's an open-source solution. Now imagine what a trillion dollar company can do.

And Google has some tool like that integrated in results to find this?

 

[howard981]

And Google has some tool like that integrated in results to find this?

Can you rephrase that, I don't understand what you are asking?

If you meant, has Google implemented AI detection, then I don't know. It's closed source.