SEO News: August 2026

Microsoft Clarity will show whether AI bots are driving traffic to your site or just stealing your content

Microsoft Clarity has introduced a new metric called “AI Scrape-to-Referral insights.” Its goal is to help website publishers assess whether AI bot activity brings them value in the form of traffic.

• Scrape/Referral Ratio: The new tab compares AI bot activity with the volume of referral traffic.
• Operator Breakdown: Allows you to identify which AI sources are driving traffic and which are only scraping.
• Detailed analysis: It leads directly to session logs for evaluating the quality of behavior of AI-referred visitors.

For website owners, this means the ability to assess more objectively which AI crawlers provide real value and which ones are simply exploiting their content.

Google is introducing new features for content personalization and the "Preferred Sources" button

Google has introduced new features for personalizing content in Search, Discover, and News, which are designed to strengthen the direct connection between publishers and readers.

• New embeddable button: Publishers can place a button on their websites that allows users to add a source to their favorites.
• Increased visibility: Preferred sources will then appear more easily in Top Stories, AI Overviews, and AI Mode.
• Over 600,000 sources: Users have already added more than 600,000 unique favorite sources.

This represents a new way to build more direct and stronger connections with readers and gain priority in Google products.

By default, Google expands the AI Overview for selected queries

Google has confirmed that, for certain search queries, it automatically expands the AI Overview to full view, eliminating the need for users to click “Show more.” The new interface also opens the “Ask anything” field by default.

• Expansion: The full AI response appears automatically, pushing organic results significantly lower.
• Goal: Google systems determine where an extension is most useful, with the aim of improving user support and engagement.
• Interaction: If the user is already scrolling, the automatic expansion stops so the user doesn't lose their position.

For publishers, this means a further reduction in the visibility of organic results in favor of AI-generated content.

Google Search Console: Track the performance of content from TikTok, Instagram, and other social media platforms

Google announced the global availability of platform properties in Search Console, which allow creators and publishers to track the performance of social content in detail across platforms such as TikTok, Instagram, X, and YouTube.

• Features: The tool is now available globally and shows how your content is performing in Search, Discover, and News.
• Detailed Reports: You can track top and trending search queries, use the 24-hour filter to quickly identify traffic peaks, and analyze the geographic origin of users.
• Implementation: For monitoring purposes, each profile (e.g., TikTok or Instagram) must be added and verified separately in Search Console, with the option to export data for a comprehensive comparison.

This data provides a valuable foundation for optimizing social media strategy and planning future content creation more effectively.

Google has completed its August 2026 spam update, which also targets AI-generated content used for manipulation

Google has completed its August 2026 spam update, which was rolled out from August 18 to 21. News from www.seroundtable.com and www.searchenginejournal.com provide further details on its focus.

• Type and Targeting: This was a “standard spam update” affecting all regions and languages, targeting websites that violate Google’s general spam policies, but not specifically targeting link spam.
• AI content: Part of the update appears to have focused on mass-generated AI content intended primarily to manipulate search results, for which Google uses the S-CTS system.
• Impact and Recovery: These actions require a review of Google's spam policies; it may take many months for the affected pages to regain visibility, as systems like SpamBrain learn gradually.

In Google's view, focusing on the actual usefulness and quality of content is far more important than mass-producing it for the sake of rankings.

Source: https://www.seroundtable.com/google-august-2026-spam-update-41895.html · https://www.seroundtable.com/google-august-2026-spam-update-done-41906.html · https://www.searchenginejournal.com/reports-indicate-googles-spam-update-focused-on-seo-ai-content/586978/

Anthropic Labels AI Content from Claude: Watermarks, Metadata, and Compliance with the EU AI Act

Anthropic is introducing new ways to label content generated by the Claude AI model to comply with the requirements of the European AI Act, as reported by aifirst.cz and support.claude.com.

• Text Watermarks: Starting August 2, 2024, invisible watermarks—consisting of a specific selection of words—will be globally embedded in all text generated by the Claude model.
• Metadata and Future Tagging: Generated files (e.g., .svg, .png, .jpg) will have digitally signed metadata in accordance with the C2PA standard, and new models launched in the EU after August 2, 2026, will contain machine-readable tags.
• Detection: The detection feature is now available in private preview for regulators and select organizations, but the public tool has not yet been released.

The origin of AI-generated content will become increasingly transparent and potentially verifiable, which will affect the credibility of such content and how it is handled.

Schema Markup and AI Citations: Debunking Myths About the Impact of Data on Generative Search

A systematic review of ten studies reveals that claims about the positive impact of schema markup on citations in AI search are often unsubstantiated; Kurt Fischman and Ahrefs found no effect—or only a slight negative effect—in their controlled studies.

• Correlation Bias: Studies showing a positive correlation lacked key control mechanisms, particularly the influence of organic search rankings.
• Rank dominance: A page’s position in organic search results—not schema—is the main predictor of whether the page will be cited by AI systems.
• Hidden data ignored: Tests by searchVIU showed that AI models (with the exception of Gemini) often do not read hidden structured data such as JSON-LD, but instead focus on visible text.

The effort spent on schema markup is now better invested in improving rankings and optimizing the visible content on the page.

Source: https://www.danielkcheung.com/musings/schema-ai-citations-evidence-review

Microsoft Clarity now distinguishes between branded and non-branded AI queries

Microsoft has enhanced the Clarity Citations Dashboard by distinguishing between branded and non-branded AI queries. The tool now provides a more detailed view of how AI cites sources in search results.

• New categorization: The Dashboard now distinguishes between AI-generated mentions related to branded and non-branded queries.

For website owners, this means a more detailed overview of how AI in search references their brand compared to general search queries.

Source: https://clarity.microsoft.com/blog/branded-non-branded-queries/

Google will allow users to disable visible watermarks on AI-generated images

Google announced that it will allow users to remove visible watermarks from their AI-generated content, including images, videos, and songs. The company is responding to the need to balance creative control and security.

• Identification: Invisible SynthID watermarks and C2PA metadata are retained for transparency.
• Availability: The option to turn off watermarks will be available in Gemini, Google Flow, and soon in Search.
• Tools: Google also makes the Credentio library available to developers for verifying content in apps.

For content creators, this means greater flexibility in integrating AI-generated media into their projects without compromising transparency.

Google Is Changing the Rules for Reviews in Structured Data

Google has updated its structured data guidelines regarding review snippets and now prohibits fake or undisclosed incentivized reviews.

• Fake reviews: They must not be based on actual experience with the product or service.
• Paid reviews: Reviews written in exchange for compensation (e.g., money), unless this is clearly and visibly stated.

Ensuring the transparency of reviews in structured data is crucial for websites to remain compliant with the guidelines.

Internal Search Results: Block or Allow Indexing?

John Mueller from Google recently addressed the question in a video of whether it is advisable to block internal search results pages from being indexed by Google. He explained that the answer is not clear-cut and depends on the specific context of the website.

• Recommendation: For most websites, it is recommended to block internal search pages.
• Main reasons: It prevents wasting the crawl budget and the creation of duplicate content.
• Exceptions: Websites with valuable, unique internal search results may benefit from indexing.

For SEO specialists, this means they must carefully assess the true value and uniqueness of internal search pages for users.

Sources

• https://searchengineland.com/microsoft-clarity-ai-scrape-to-referral-insights-report-484959
• https://blog.google/products-and-platforms/products/search/personalize-search-discover-news/
• https://www.seroundtable.com/google-ai-overviews-push-ai-mode-responses-41974.html
• https://developers.google.com/search/blog/2026/07/platform-properties-social-video-guide
• https://developers.google.com/search/docs/monitor-debug/analyze-social-video-content
• https://www.seroundtable.com/google-august-2026-spam-update-41895.html
• https://www.seroundtable.com/google-august-2026-spam-update-done-41906.html
• https://www.searchenginejournal.com/reports-indicate-googles-spam-update-focused-on-seo-ai-content/586978/
• https://www.aifirst.cz/clanky/claude-pise-do-textu-neviditelny-vodoznak
• https://support.claude.com/en/articles/16266773-how-claude-marks-ai-generated-content
• https://www.danielkcheung.com/musings/schema-ai-citations-evidence-review
• https://clarity.microsoft.com/blog/branded-non-branded-queries/
• https://techcrunch.com/2026/08/14/google-will-now-allow-users-to-remove-visible-watermark-from-its-ai-generations/
• https://www.seroundtable.com/google-review-snippets-guidelines-update-41754.html
• https://www.youtube.com/watch?v=cfIQ_ksB61U