Show notes
E1216: Anthropic is expanding AI text watermarking to more Claude models on November 3. Other AI companies have also committed to watermarking their outputs. What does this mean for SEO? Will Google start penalizing AI-generated pages, and can it tell the difference between content written entirely by AI and content edited with AI? I break down the risks of publishing unedited AI content, what Google's recent guidance suggests, and why I still expect carefully reviewed, AI-assisted pages to rank and even be trained upon. In this episode: - Claude's expanding text watermarking: Which models Anthropic says will next receive watermarks why this matters for publishers. - The Ahrefs AI content study: 5.3% of top-ranking pages were classified as fully AI-generated, another 9% as at least 80% AI-generated, and 80% of high-AI pages were indexed compared with 49% of low-AI pages. - Google's position on AI content: Quality rater guidance for unoriginal, low-effort pages, warnings about hallucinations, and the risks of publishing AI pages at scale without proofreading. - The possibility of tougher Google updates: How some SEOs expect more action against scaled AI content, and why that remains a prediction rather than a confirmed update. - My GPTZero test: A roughly 400-word SEO landing page was identified as human-written and AI-polished, while an AI-generated article about me was classified as 100% AI. - Real examples of AI hallucinations: False claims I found in an AI-generated article about my own work, videos, and audience. - AI detection as a spectrum: Why assisted writing, AI polishing, and fully generated text shouldn't be treated as identical. - How Google might evaluate AI-heavy sites: My theories about user behavior, sitewide patterns, branded searches, and future AI detection. These are possibilities, not confirmed ranking factors. - AI content in model training: The concern that training on unverified AI-generated material could increase errors and hallucinations. - How we're writing pages for a SaaS I'm funding: Using AI line by line as a writing assistant while manually reviewing the language ourselves. - Building branded searches through short-form video: How SEO entrepreneur Ilias ISM hired creators for $500 per creator per month to post daily videos, generate millions of views, and increase searches for his brand. My main point: AI watermarking doesn't mean every page touched by AI is destined to lose rankings. There's an important difference between publishing hundreds of unchecked AI pages and using AI as a writing assistant. ⭐️ Claude Watermark Announcement - https://pasqualepillitteri.it/en/news/27785/claude-watermark-november-3-opus-4-7-sonnet-4-6 ⭐️ Ep 1210 - The Backlinks Worth Paying $2,000 For (w/ Ilias ISM) - https://www.youtube.com/watch?v=5NRS4ibsMKg ⭐️ My Exact Social Media Posting Strategy - https://edwardsturm.com/articles/social-media-posting-strategy/ 💎 Compact Keywords - My SEO Course - Get paying customers through SEO - Clear step-by-step video breakdowns - SEO templates to be copied and adapted for your products and services: https://compactkeywords.com/ 00:00 Anthropic Watermark Email 00:35 AI Content Panic 00:50 Ahrefs AI SEO Stats 01:31 Google Cracks Down 02:45 Hallucination Example 03:41 Detection Is a Scale 05:42 How Google Penalizes 07:27 Training Data Concerns 08:34 Branded Search Signals 08:59 Buy Branded Searches 11:13 Wrap Up The Edward Show. Your daily search engine optimization podcast: https://edwardsturm.com/the-edward-show/ #searchengineoptimization #seo #digitalmarketing #seopodcast