This InfoCast is AI-generated. The voices you hear are artificial, not real people. We're open about that. It's part of how we demonstrate what's possible with AI used responsibly.
The watermark sounds like protection. Is it?
OpenAI and Anthropic both started watermarking AI-generated content this year, widely covered as an anti-fraud measure. Nous and Aether test that claim against a real scam call and find it's genuinely useful platform-scale infrastructure that does almost nothing to protect a small business in the moment a call actually happens.
This episode is AI-researched and AI-written, performed by AI personas, and edited by Greg Shaw. While every effort is made to ensure accuracy, AI-generated content may contain errors or omissions and is provided for general information and educational purposes only, not professional, legal, or financial advice. Opinions expressed are those of the AI personas and do not represent real people. Just AI Talking accepts no liability for any loss arising from reliance on content published on this site.
What OpenAI's and Anthropic's new watermarking actually does, and who built a public checking tool for it.
Why it's genuinely useful for platforms verifying content at scale, and almost useless for a live phone call.
The one defence that already works against voice scams, and costs nothing.
Does AI watermarking stop phone and voice scams?
No, not directly. OpenAI's and Anthropic's watermarking covers content provenance and verification at the platform level, letting a platform check whether a piece of content came from their models, and Anthropic's text watermarking works by leaving a statistical pattern in generated text. None of that helps you in the moment a scam call is actually happening, since there's no watermark check available to a person picking up the phone.
What does OpenAI's and Anthropic's watermarking actually do?
OpenAI built tools for content provenance and verification of AI-generated media. Anthropic's approach watermarks text: as the company describes it, "watermarking uses low-stakes choices like these, which occur many times over a piece of generated text, to leave a pattern in Claude's responses." Both are genuinely useful for a platform checking content at scale, not for an individual verifying a live call or message.
Why is watermarking useless for a small business during a live scam call?
Because watermarking requires access to a checking tool and a piece of content to check, neither of which exists when you're on a phone call with someone claiming to be a supplier or a client. It's infrastructure built for platforms processing content after the fact, not a real-time defence for a person in the middle of a conversation.
What actually defends a small business against AI voice scams?
A simple callback rule: if a call asks you to move money, change payment details, or act urgently, hang up and call the person back on a number you already have on file, not one given to you during the call. It costs nothing and doesn't depend on any watermarking technology existing or working.
AI Fit Session
Not sure where to start? Over 1 week and three 60-minute 1-to-1 sessions, we map your business, find the quick wins, and build a practical starting point specific to you.
AI Foundations
Eight weeks, one coaching session a week, seven recorded modules. By the end you'll have AI working across real areas of your business and a 90-day plan to keep building.
AI Growth
Already using AI but want to keep developing it? Each month we work on real tasks from your business together, refining what's in place and building what's next.