If an AI watermark scares you, your content may be the real issue

If you were anyplace adjacent the net past week, you apt heard that Anthropic announced machine-readable watermarks for Claude-generated text.

Come on! Why do you really attraction astir watermarks, aliases honestly, immoderate different awesome indicating that AI touched a portion of content?

Are you worried that group will observe you’re publishing crap? Or is the existent “issue” that you deliberation hunt engines and AI reply engines will penalize you for utilizing it?

You should perfectly beryllium concerned astir the first question.

But let’s talk why truthful galore group are really panicking:

“Da rankingz, Nick!”

You’re not acrophobic of the watermark. You’re acrophobic of being recovered out.

For galore people, the fearfulness isn’t the watermark itself. It’s what others mightiness presume aft seeing it: that they deficiency expertise, didn’t do the work, aliases charged for an thought a instrumentality produced.

You can’t power those assumptions. You tin only power what you publish. (Don’t worry, group are moving difficult to not get caught)

If being “found out” would embarrass you, the problem astir apt isn’t the watermark. It’s that you don’t judge the vanished activity tin withstand scrutiny.

There’s besides a monolithic quality betwixt utilizing AI to shape your ideas, situation your thinking, aliases amended a draught and blindly publishing its first response.

A watermark whitethorn bespeak that Claude participated successful the process somewhere. It can’t show america who originated the idea, really overmuch AI contributed, aliases whether the vanished activity is meticulous aliases valuable.

Evidently, anti-watermarks are a business now, too. I share, for example, but I cognize astatine slightest 1 of you is going to test/purchase it.

A watermark mightiness place portion of the accumulation process. It can’t measure the vanished product.

But estimation isn’t the only concern. Marketers are already asking whether hunt engines will dainty that awesome arsenic a logic to suppress the content.

Google doesn’t attraction that you utilized AI. It cares what you did pinch it.

Despite what immoderate folks desperately want to believe, Google has ne'er said that AI-generated contented is automatically penalized aliases moreover “bad”.

In fact, Google’s ain guidance acknowledges that generative AI tin beryllium useful for researching topics and adding building to original content.

The rumor present is erstwhile businesses usage automation to create pages astatine standard without adding immoderate meaningful value. The net has capable copycats; if you’re letting AI make it solely connected its own, it’s going to transcript each the copycats…

Google’s scaled contented maltreatment policy aims to target ample volumes of low-value contented created to instrumentality Google into achieving higher hunt rankings. This isn’t new; earlier AI, Google fought garbage content. AI conscionable injected the spam pinch steroids.

AI contented isn’t the problem. Unoriginal contented created chiefly to manipulate rankings is the problem. (I’m frothing astatine the rima for the AI-Penguin type rollout, crushing those who laid disconnected their teams and pushed 100% AI contented live.)

(Funny enough, Google just completed the August 2026 spam update!)

When fixed afloat autonomy, AI makes it easier to nutrient generic advice, unsupported claims, near-duplicate pages, recycled hunt results, and thousands of pages cipher asked for… fto unsocial spot worth in.

Yes, creating garbage tin decidedly wounded your SEO. The watermark itself is not what causes this.

GEO isn’t an ai purity title either

There is nary cosmopolitan norm that “Claude watermarked contented will beryllium excluded from AI-generated answers.”

Answer engines still request accusation that is useful to people.

The applicable mobility isn’t:

“Can the exemplary show that AI helped create this?”

It’s:

“Does this page lend thing worthy retrieving, citing, aliases synthesizing?”

Author bylines are a communal starting constituent to build consistency for AI bots. However, by itself, it can’t make repetitive accusation “citation-worthy.” On the different extremity of the spectrum, an AI watermark doesn’t erase the worth of original research, firsthand experience, aliases moreover opinions that make quality interest.

“A quality wrote it” isn’t a competitory advantage if a 1000 different humans already wrote the aforesaid thing. (I won’t moreover commencement connected AI and copyright infringement.)

Go ahead. Scan this post.

Substack will now fto you scan posts connected its level utilizing Pangram integration and get an thought of conscionable really “AI” that contented is. (They stock really it useful here.)

So, spell ahead. Scan this post. Here I’ll screenshot it myself.

8%, 18%, possibly it says 80% AI. (OMGz!)

I genuinely don’t care. (Although 31% feels very precocious for really galore keystrokes I’ve added to constitute this post, but meh!)

I ne'er promised that each condemnation of this article was written successful the wilderness pinch nary net connection. You publication it because the title caught your attention, you’ve weighted my erstwhile work, aliases simply get intermezo from my spicy takes connected each things search/AI/business.

An AI people cannot show maine whether you vanished the post, shared it, subscribed, clicked my sponsor’s link, aliases hired me.

“Eighteen percent AI” is simply a random number. Attention, trust, subscribers, and gross are winning.

If this station produces those outcomes, it did its job. If it’s useless, a “100% human” people doesn’t instantly springiness maine 10,000 much subs (bummer!)

Care astir what you’re consenting to publish

The rules ne'er changed. Garbage, written by a quality aliases AI, isn’t support to standard to a gazillion pages. It makes your name/brand look for illustration garbage, and you suffer trust.

AI doesn’t negate your work successful providing value work.

Before publishing, inquire whether the activity is accurate, original capable to warrant existing, useful to its intended audience, and thing you’re consenting to take sides pinch your sanction attached. (Now you understand why group are building the “remove whitelisting products)

If it is, people it and guidelines down it.

If it isn’t, manipulating the AI people won’t make it immoderate better. (Come on, we each learned successful our SEO 101 days astir “duplicate content,” and the rules aren’t immoderate different for AI.)

You request to attraction connected winning.

I mean, don’t want the watermark? You could [gasp] constitute the full contented connected your own!?

I usage AI to make #SEOForLunch leader images and thief edit these articles. Otherwise, my rants would routinely transcend 1,500 words.

So, really, you should convey AI for making maine unopen up sooner.

This station first appeared connected the author’s website and is republished present pinch permission.