Fake content is getting easier to spot

Hey there Smarty 👋

Are you geared up to catch the latest and greatest in quick shorts?

And just a quick heads-up! If you stumbled upon us through a friend, make sure to subscribe here! That way, you’ll never miss out on the trending shorts.

🔏 As AI content gets easier to spot, what's left becomes worth more

Watermarking and detection technology for AI-generated content is tightening across the industry, with distribution controls increasingly able to flag synthetic content at the point of publishing rather than relying on a reader to guess. 

That's usually framed as a compliance story, one more thing to disclose, one more label to apply. It's also a market story: as detection gets more reliable, the content that clearly wasn't AI-generated becomes a distinguishable, verifiable category of its own for the first time, rather than an assumption nobody could confirm either way.

A market where anyone can plausibly claim their content is authentic is one where the claim is worth very little. 

A market where authenticity can actually be verified is one where it becomes a real signal, both to a human reader deciding what to trust and increasingly to the AI systems that weigh source credibility when deciding what to cite.

Audit which of your existing content can actually be verified as human-made

Most teams can't currently answer, with any confidence, what percentage of their published content involved generative tools somewhere in production, since the line between human-written and AI-assisted has blurred inside normal editing workflows for a while now.

Pull your highest-value published content and trace its actual production path: fully human, human with AI-assisted editing, or generated outright. 

This isn't a compliance exercise. It's an inventory of which assets could credibly carry a verified-authentic label once the market starts rewarding one.

Build provenance tracking into new content from here forward

Reconstructing production history for content already published is difficult once the people who made it have moved on or forgotten the details. Capturing it going forward costs almost nothing.

Add a single field to your content workflow at creation time: human-authored, AI-assisted, or AI-generated, with enough detail to stand behind the label later. 

That record is the raw material for whatever verification standard the market settles on next, and it's the kind of provenance Semrush's AI Visibility Toolkit tracks alongside how your brand is cited across ChatGPT, Perplexity, Gemini, and Google AI Mode.

Treat verified authenticity as a citation signal, not just a trust signal

As detection tools make synthetic content easier to flag, AI systems weighing which sources to trust and cite have a new input available that didn't reliably exist before: whether a source's content is independently verifiable as human-made.

Authenticity used to be a claim anyone could make. It's becoming something you can actually prove, and proof is worth more than a claim. You can try it free for 7 days.

Together with AirOps

How Stripe and OpenAI Built Marketing From Scratch

Higher growth targets are landing on teams that already feel stretched. Without sharper priorities and reporting, every new request creates pressure without improving performance.

On September 17 join AirOps CMO Christy Roach and Krithika Shankarraman to show how faster-growing marketing teams operate, using findings from 300+ marketing leaders.

You’ll learn how to:

• Prioritize your biggest growth constraints without letting reactive work take over
• Report performance clearly and build leadership confidence in channels like AI Search
• Align with the wider business so promising marketing investments move faster

Krithika was the first marketing hire at Stripe and OpenAI. Christy’s experience spans AirOps, AssemblyAI, Airtable, Asana, and Gusto.

You’ll walk away with operating principles you can immediately use to tighten priorities and make clearer investment decisions.

🔎 OpenAI is building more of search for itself

OpenAI appears to be reducing its dependence on traditional search engines just as ChatGPT becomes a bigger destination for finding information. The bigger shift is what happens to the websites supplying those answers.

The Breakdown:

OpenAI is testing its own search index: Researchers found an experiment named “prefer-index-over-serp-v3,” suggesting OpenAI is comparing results from its own index against third-party search engine results.

The search infrastructure is getting broader: Separate indexes were identified for news, shopping, PDFs, YouTube, and academic papers, alongside a reported experiment involving crawling as many as one billion webpages.

ChatGPT is becoming a major destination: ChatGPT.com reached 1.09 billion monthly US visits, up 48% year over year and ranking ninth nationally, while Bing traffic fell 50% and DuckDuckGo declined 16%.

AI search means fewer clicks outward: Although 95% of ChatGPT users still use Google, forcing users into AI Mode reduced external click-through by 18.8 percentage points, with even steeper declines for Reddit.

OpenAI building its own index would give it more control over how information is discovered and retrieved. At the same time, AI interfaces are becoming destinations themselves, creating a growing gap between being used as a source and actually receiving the click.

👨‍💻 Quick Hits

📈 Acquco grew monthly affiliate sales from $1,500 to $36,000 in four months after adding 130+ creator partnerships. Levanta’s self-diagnostic identifies roster gaps that prevent your program from finding and activating its next wave of revenue-producing partners. Take the Self-Diagnostic.

💡 Google rolled out an AI-powered AdSense Help assistant to all English-language traffic, letting publishers ask conversational questions about onboarding, features and troubleshooting, though answers may occasionally be inaccurate.

🤖 Harvard research found little resistance to AI replacing search marketers, with the occupation scoring 2.31 out of 7 for moral objection. Support increased significantly when AI outperformed humans.

🛍️ Google Merchant Center accidentally exposed internal AI prompts behind its performance insights, revealing how the system analyzes product changes, selects highlights, follows formatting rules and checks responses before displaying them.

🛒 YouTube is testing localized affiliate product tags that automatically connect creator-tagged products to eligible local retailers, potentially improving shopping performance and reducing delivery times while supporting nearby merchants.

That’s a wrap for today! Tell us your thoughts about today's content as we line up more Shorts! And don’t hesitate to share this with someone who’d adore it. 🥰