Google Drops the Visible Tag, Keeps the Hidden One
Google is letting users turn off the visible watermark on AI generations, according to TechCrunch. The setting doesn't touch the invisible benchmark marking Google uses to detect AI-generated files — that stays regardless of what the user chooses. A Gizmodo piece circulating on Hacker News read the move as opening the door wider for unlabeled AI content.
The setting
The facts are narrow, and TechCrunch reports them plainly: Google will now let users remove the visible watermark it applies to AI-generated images and video. Flip the setting, and the on-image mark that told a viewer "this was made by AI" is gone.
But the same report is careful to note what the setting doesn't do. Turning it off has no effect on the invisible benchmark markings Google uses to identify a file as AI-generated. Those persist. The user controls what a human sees. Google keeps what its own systems can read.
Removability was never symmetric. The mark a stranger could see on a suspicious image is now optional. The mark Google's own detection tools rely on isn't. That's not a bug in the design — it's the design.
The reaction
Not everyone read the change the way Google's product notes probably frame it. A Gizmodo piece, surfaced on Hacker News under the headline "Google Opens the Gates of AI Slop Hell," treated the toggle as exactly what its title says — a gate, opened wider. The Hacker News thread logged five points and one comment, a small case file by itself, but the framing is the tell: strip the visible marker and what circulates online loses its label, even if the file itself still carries an invisible one only Google can read.
A watermark only protects a public that can see it. An invisible benchmark protects whoever holds the detector. Those are two different audiences, and this change serves only one of them.
What stays behind
TechCrunch's reporting doesn't say who else besides Google can read the invisible marking, or under what circumstances it gets checked. That's the part of the file that's still open. What's confirmed is narrower and, in its way, more telling: the identifier a casual viewer relies on is now a preference. The identifier Google's own systems rely on is not up for a vote.
Questions people ask
Does turning off Google's visible watermark remove all evidence an image is AI-generated?
No. TechCrunch reports the setting only removes the on-image visible mark. The invisible benchmark marking Google uses to identify AI-generated files is unaffected by the toggle.
Who controls whether a file's invisible AI marking gets removed?
Not the user. The invisible marking stays regardless of the visible-watermark setting, per TechCrunch's reporting.
Why did commentators call this "AI slop hell"?
A Gizmodo article used that phrase to describe the change, arguing it makes AI-generated content harder for viewers to identify at a glance. The piece was aggregated on Hacker News with 5 points and 1 comment.
Is this the same as removing all watermarking from Google's AI tools?
No — it's specifically the visible, on-image watermark that becomes optional. The invisible benchmark identifier is a separate mechanism and is not affected.
Comments
Post a Comment