Anthropic, the maker of the chatbot Claude, kicked off a firestorm by announcing it would start watermarking text produced by the bot.
CFOTO/Future Publishing via Getty Images
Two hundred twenty-five years after Alexander Hamilton founded the New York Post, an AI chatbot bearing his name is now helping readers decide what to read there.
Open the tabloid’s mobile app, and there it is at the bottom. Tap on the Hamilton icon, and the user is greeted by name with a new screen offering “more ways to explore the stories that matter to you.” That promise includes an interactive summary of the day’s news, along with a personalized digest of stories tailored to the user’s interests.
“For nearly 225 years, the Post has evolved alongside the way people consume news, from print presses to smartphones and now to AI,” Sean Giancola, CEO of the New York Post Media Group, said in an announcement about the new tool.
The bot named in honor of Hamilton, he continued, is “the next evolution” of that journey.
Media companies going all-in on AI
Across the media business, variations on the Post’s same AI experiment are becoming increasingly common. Netflix, for one, is now using generative AI across hundreds of titles. Time built an AI agent around its archive, while The Economist launched inside an app ChatGPT—and USA Today embedded an AI answer engine that lets readers ask conversational questions about its journalism.
Roku in recent days added an AI channel to its free channel lineup, with this new offering the first fully ad-supported channel featuring AI-generated programming around the clock. Even the commercials are AI-generated.
The Arena Group, owner of Parade, Men’s Journal and The Street, is going further still. The company rebranded itself as Paradium.AI, it acquired AI content generator InfoSentience and launched its own AI-assisted platform for generating articles and video.
This all comes as companies like Google, Meta and OpenAI increasingly want users to become AI-powered media producers themselves, generating images, video, and text-based content with a simple prompt.
But even as media companies make it easier and more normal to interact with AI tools like the Post’s Hamilton, there’s another movement unfolding in the opposite direction.
Call it AI’s scarlet letter era, as media companies and platforms on the one hand encourage widespread use of the technology—while, at the same time, there’s an increasing push to label and even disadvantage the content AI produces.
The rise of AI watermarks and content labels
Spotify offers one example.
The music streamer has just announced a push to start labeling artist profiles built around artificial identities with an “AI Persona” badge. The labels, set to roll out in mid-September, are meant to distinguish entirely AI-generated artists from real musicians who use AI as part of their creative process.
Spotify is also going a step further. Music tied to “AI Persona” profiles will be excluded by default from its editorial and algorithmic recommendations.
Speaking of music, the AI-music generation platform Suno plans to introduce audio watermarking. YouTube is also now adding an AI label to otherwise realistic-looking content, even when a creator hasn’t disclosed the use of AI, while TikTok says billions of videos have already been marked as AI-generated there.
If all this wasn’t confusing enough, Google announced on Friday that it will now let users remove the visible watermark from AI-generated images, videos and music created with several of its models. The company isn’t abandoning AI IDs altogether—invisible watermarking, for example, won’t go anywhere—but the move adds yet another wrinkle to the emerging rules around labeling AI.
And then there’s Anthropic, which is adding an invisible watermark to text produced by its newest Claude models. That change is largely a response to new EU transparency rules, but Anthropic is rolling it out everywhere Claude is used.
The news about Claude watermarking kicked off a firestorm on X, with users panicking that AI-assisted writing will now bear a kind of permanent “scarlet letter.”
“I had ditched Grammarly for Claude for proofreading because it does a better job,” blogger Erick Erickson said in a post on X. “But now the stuff I’ve written will be watermarked that Claude did the work. This is ridiculous.”
Between the proliferation of AI usage and now the corresponding watermarks and labeling, it can feel like there’s almost a kind of cognitive dissonance at work—in that media companies are conditioning more people to use AI, while at the same time platforms, in particular, are training those same users to be suspicious of what AI produces. Normalizing AI, in other words, while at the same time stigmatizing the output.
And there’s another concern, spelled out in a July 2026 paper by researcher Federico Germani. He writes, about watermarking AI content: “Invisible watermarking encodes only model origin; when operationalized into visible AI-generated labels, it reduces complex creative processes to a misleading binary and provides no information about truthfulness. Such labels may stigmatize legitimate uses of generative tools while encouraging misplaced trust in unmarked content.”
That’s the destination all this may be heading toward.
AI is becoming something audiences are encouraged to use constantly, but also something they’re being trained to be wary of. And therein lies a certain irony: AI watermarks and labels are supposed to help people know what to trust, but they may end up making unlabeled content seem more trustworthy than it deserves to be.
Leave a comment