Gemini's SynthID actually impact indexing?

[MsLin]

Google's SynthID watermarking for AI text. Is there any way we can remove it? rewriting or paraphrasing kill the watermark? What actual negative impacts have you guys faced from this?

 

[harpage]

Google's SynthID watermarking for AI text

I just googled this now and was surprised that I completely missed that AI generated content could be watermarked.

If the content is slightly modified, does that break the id chunk? How much modification is needed.

 

[MsLin]

I just googled this now and was surprised that I completely missed that AI generated content could be watermarked.

If the content is slightly modified, does that break the id chunk? How much modification is needed.

Yeah, that is exactly what I want to know. If we edit the text, does the watermark go away? And how much edit is enough? Hope someone can answer

 

[HotMilfNearYou]

Yeah, I discovered this thing by accident. Gemini didn't understand which feature I wanted explained, so it just gave me a full breakdown of it instead. He never once mentioned whether it affects SEO or not. In my opinion, Google isn't going to reveal it for avoid it will be too easy for us. We'll have to wait for experts to run tests and publish their results. That being said:

I prefer to assume it penalizes ranking. Because even if it doesn't, it's only a matter of time before they change their minds. They always end up becoming stricter, but we don't know when.

Imagine you manage to rank content that suddenly drops in rankings because of this. That's not cool.

 

[Fourthlevel]

That's really interesting, didn't know much about that before today. Thanks for sharing

 

[bulkcloudaccs]

Google says AI writing is not the problem; the problem is low-quality/spammy content, not Gemini's SynthID ranking signal.
SynthID cannot be removed outright, but deep human rewrites and value additions do not appear to have any real negative impact.

 

[medialounge]

I just googled this now and was surprised that I completely missed that AI generated content could be watermarked.

If the content is slightly modified, does that break the id chunk? How much modification is needed.

Every word done with AI is actually watermarked
Currently testing with quilbot to see if it's able to deal 100% with it

 

[harpage]

Every word done with AI is actually watermarked
Currently testing with quilbot to see if it's able to deal 100% with it

Interesting. How are we sure that Quilbot is not using content with SynthiD to "humanise" AI written content?

 

[RapidWorkers]

Google's SynthID watermarking for AI text. Is there any way we can remove it? rewriting or paraphrasing kill the watermark? What actual negative impacts have you guys faced from this?

Detection tools can flag watermarked text but there’s no clear evidence that having watermarked content causes direct penalties in search or elsewhere yet

 

[liomarver]

Honestly, this feels a bit overhyped to me. I haven’t seen any real proof that SynthID alone hurts rankings. Google keeps changing the narrative, but at the end of the day, good content still ranks

 

[MsLin]

Detection tools can flag watermarked text but there’s no clear evidence that having watermarked content causes direct penalties in search or elsewhere yet

Let us know how it goes, curious if Quillbot can actually kill the watermark

 

[MsLin]

Detection tools can flag watermarked text but there’s no clear evidence that having watermarked content causes direct penalties in search or elsewhere yet

My concern is mainly about future updates