Google AI Mode running on Gemini 3.8 Flash is producing detailed answers to top-of-funnel queries without citations or outbound links. The responses are not short summaries. They often contain sections, tables, bullet points, explanations, and suggested follow-up topics.
Barry Schwartz at Search Engine Roundtable reported the pattern in August 2026. Early testing among Google’s Pro and Ultra subscribers showed fewer outbound links in AI-generated answers, especially for informational queries.
Screenshots from AI Mode show the same behavior across searches for “seo,” “marketing,” “digital marketing,” “email marketing for small businesses,” and “machine learning.” Each answer provides enough information to satisfy much of the user’s intent inside Google.
However, none of those examples includes a visible external citation or source link.
That creates an important problem for publishers. Google has previously discussed making links and citations more prominent in AI-powered search. Gemini 3.8 Flash currently appears to move in the opposite direction.
This guide explains what the screenshots show, why the change matters, and what publishers can do if citation-free AI answers become more common.
If you want to compare AI referral traffic trends with other publishers and SEO practitioners, join Scale Xpert’s Discord community for SEO learning and genuine backlink exchange.
What the Gemini 3.8 Flash Screenshots Actually Show
The screenshots reveal a consistent pattern across several informational searches. Google AI Mode provides a complete-looking answer, but the user receives no obvious path to an external source.
The “SEO” Query
For the query “seo,” AI Mode starts by defining Search Engine Optimisation. It then explains the major areas of SEO, including on-page SEO, technical SEO, and off-page SEO.
The response also includes a “Why SEO Matters” section. It discusses sustainable traffic, user trust, and user experience. Suggested follow-up topics appear after the main answer.
Despite the amount of information provided, the response contains no visible external links, citations, or source attribution.
The “Marketing” Query
The “marketing” query produces an even more structured answer. AI Mode explains the four Ps of the marketing mix and describes each element in detail.
It also displays a table covering modern marketing categories, key channels, and their main objectives. A customer journey funnel explanation follows the table.
Once again, the answer does not provide citations or outbound links.
Email Marketing for Small Businesses
The query “email marketing for small businesses” creates another useful example. AI Mode compares Mailchimp, MailerLite, Klaviyo, ConvertKit, and Brevo.
The comparison includes use cases, strengths, and starting price information. Additional sections explain legal list building, automated email flows, and open-rate best practices.
The platform names appear as plain text. Users cannot click through to the companies or verify the information through a cited source.
Digital Marketing and Machine Learning
The “digital marketing” query creates a structured table with six marketing channels. It also explains their focus and typical goals.
A search for “machine learning” follows the same format. AI Mode defines the topic, explains how machine learning works, and compares supervised, unsupervised, and reinforcement learning.
It also gives everyday examples. Yet no source links appear.
A Consistent Zero-Citation Pattern
Across all five examples, the structure changes but the citation behavior does not. Users receive rich and readable information without a visible mechanism for checking the original sources.
This matters because the answers can satisfy broad informational intent on their own. A user may have little reason to continue to another website.
The Contradiction With Google’s Public Statements
The current citation behavior appears difficult to reconcile with Google’s previous statements about AI search.
Robbie Stein, Google’s Vice President of Search, has discussed Google’s intention to increase citations and make links more prominent within AI-generated search experiences.
Those statements addressed a major publisher concern. If AI answers reduce clicks, publishers at least want meaningful attribution and opportunities for referral traffic.
The Gemini 3.8 Flash examples show something different. Answers have become detailed enough to satisfy users while the visible source layer has disappeared.
Barry Schwartz highlighted the contradiction in his discussion of the rollout. The key question now is whether the missing citations result from a temporary problem or a deliberate product decision.
Bug or Design Choice?
The distinction matters for publisher strategy.
If Google is dealing with a bug, a later model or interface update could restore citations. Publishers may then see the current behavior as a temporary deployment issue.
A deliberate design decision would carry much larger consequences. Google could continue using information from the wider web while reducing the traffic pathways that send users back to publishers.
At the time covered by this report, Google had not publicly clarified the reason for the citation reduction.
The rollout also remains important to watch because the behavior was initially associated with paid Pro and Ultra users. A wider rollout would expose many more searches to the same experience.
For additional context, read our guide on how Google AI Overviews select sources and how to get your content cited. It explains how Google’s source-selection behavior has traditionally worked and why the Gemini 3.8 Flash examples stand out.
Why This Matters More Than a Normal AI Overview CTR Drop
Publishers already know that AI-generated search answers can reduce organic click-through rates.
Previous research around AI Overviews suggested substantial declines in clicks to traditional results. Ahrefs reported major reductions for pages affected by AI Overviews, while Seer Interactive also found significant declines for some non-branded queries.
Those situations still left publishers with one important opportunity: citations.
A user could read the AI summary and then click a cited page for more information. That created at least a partial referral channel.
Gemini 3.8 Flash introduces a more extreme scenario. When an AI response contains no citation, the citation-based referral opportunity disappears entirely.
AI Mode Changes the Click Model
AI Mode also differs from a standard AI Overview because it creates an immersive search experience.
Users can ask follow-up questions and continue the conversation inside Google. Each new answer can provide more context without requiring a visit to another website.
Traditional search works differently. A title and snippet usually provide only part of the information. Users often need to click an organic result to complete their research.
AI Mode aims to answer much more of the query inside the interface itself.
This makes citation behavior especially important. A complete answer with visible citations can still create discovery and referral opportunities. A complete answer without citations provides publishers with far less direct benefit.
Publishers considering their options should also review the decision framework for whether publishers should opt out of Google AI search features.
However, publishers should not assume that every control used for AI Overviews works in exactly the same way for AI Mode.
The Google-Extended and Robots.txt Situation for AI Mode
The citation issue raises another question: what control do publishers have over content used by Google’s AI products?
Google-Extended is a robots.txt token associated with certain Google generative AI uses. Some publishers use it because they want more control over how Google uses their content in AI systems.
A typical implementation looks like this:
User-agent: Google-Extended
Disallow: /The difficult part is understanding exactly how those controls interact with AI Mode.
AI Mode operates as a search experience while also using Gemini models. That creates uncertainty about which crawling, indexing, retrieval, and AI-use controls apply to each part of the system.
Publishers should therefore avoid assuming that one robots.txt directive automatically controls every Google AI feature.
Our guide to Google AI opt-out controls, Search Console, and Google-Extended explains the available controls in more detail and shows where the remaining limitations exist.
What Publishers Can Actually Do About Zero-Citation AI Mode
The current evidence does not justify panic. The behavior still needs monitoring, especially while the rollout remains limited.
Publishers can still take several practical steps now.
1. Measure Before You React
Start with your own data.
Check whether traffic from AI platforms represents a meaningful source of sessions for your site. Then watch whether those sessions change as Google adjusts AI Mode.
Set up AI Assistant channel tracking in Google Analytics 4 so you can separate AI-referred sessions from other acquisition channels.
Many publishers still receive relatively little measurable referral traffic from AI Mode. If that describes your site, a citation change may not yet have created a material traffic loss.
Measurement helps you separate an industry-wide concern from an actual problem in your analytics.
2. Build Brand Visibility Even When Links Disappear
Links are valuable, but they are not the only type of visibility that matters.
AI systems frequently mention company names, tools, publishers, products, and experts directly inside their answers.
For example, the email marketing response mentioned Mailchimp and other platforms without linking to them. A user may still remember one of those brands and search for it later.
That means brand recognition becomes increasingly important in a low-link environment.
Publishers should strengthen the signals that connect their brand with specific topics, problems, products, and areas of expertise.
Our four-tier framework for brand authority signals that help AI systems recommend your brand explains how those signals work.
3. Diversify Beyond Organic Search Clicks
Publishers that depend almost entirely on Google organic traffic face the greatest exposure to AI search changes.
Owned audiences provide more protection.
Email newsletters can bring readers back without requiring another Google search. YouTube subscribers, podcast audiences, communities, direct traffic, and app users create additional paths to your content.
Diversification will not replace SEO. It reduces the damage when one acquisition channel changes quickly.
The creator traffic survival strategy for the Google AI search era covers practical ways to build those additional channels.
4. Prioritize Bottom-of-Funnel Content
The current Gemini 3.8 Flash examples focus heavily on broad informational searches.
Queries such as “seo,” “marketing,” and “machine learning” have simple educational intent. An AI system can often answer them with general information.
Bottom-of-funnel searches behave differently.
A person comparing specific products, services, tools, or solutions often needs details that lead to an action. These searches may require more evidence, current information, product pages, comparisons, or specialist sources.
Publishers should review their content portfolio and identify pages that solve specific decision-making problems.
Our analysis of AI Overview CTR impact by query type and what publishers need to know provides additional context for deciding which topics deserve investment.
The Google News Tab Disappearance Bug: A Related Publisher Issue
The same Search Engine Roundtable discussion also highlighted a separate issue affecting some news publishers.
Fresh articles may appear in Google’s News tab shortly after indexing and then disappear around 20 to 30 minutes later.
The reported pattern follows a simple sequence. A publisher releases an article. Google indexes it successfully. The page initially appears in standard search and the News tab.
After a short period, the article disappears from News. It may return several hours later.
Standard “All” and “Web” results do not necessarily show the same problem.
Why the First 30 Minutes Matter
The timing creates a serious concern for publishers that cover fast-moving stories.
Breaking news often receives much of its early attention immediately after an event occurs. Readers search while the topic is fresh, and publishers compete for visibility during that initial window.
An article that disappears from the News tab during that period can miss an important opportunity.
Early traffic can also generate secondary benefits. Readers may share the story, reference it on social platforms, or link to it from other publications.
A temporary visibility problem can therefore have effects beyond the missing News clicks.
Publishers that rely heavily on Google News should monitor fresh articles during the first several hours after publication. Record when each article enters and leaves the News tab so you can identify a consistent pattern.
What Gemini 3.8 Flash Signals About the Direction of AI Search
Even if Google later restores citations, the current examples reveal an important capability shift.
AI Mode can already generate useful answers to broad educational queries without requiring users to open another page.
A search for “seo” can return a definition, explain the main SEO categories, describe the benefits, and suggest additional questions. From the user’s perspective, that may be enough.
Missing links do not necessarily make the answer feel incomplete.
That changes the role of generic informational content.
Publishers have historically attracted search traffic by explaining common concepts better than competing pages. AI systems can now synthesize many of those explanations directly in search.
Why Unique Information Still Matters
Commodity information faces the highest risk because many websites contain similar facts and explanations.
Original information creates stronger differentiation.
Examples include proprietary datasets, surveys, original experiments, unique case studies, primary reporting, expert commentary, and methodologies developed by your own team.
An AI system can easily summarize a common definition from information available across thousands of pages. It has a harder time replacing research that exists on only one source.
This does not guarantee a citation. However, unique information gives publishers a stronger reason to become the identifiable source behind a claim.
Our guide on why deep content matters more in the AI search era explains this shift in more detail.
You can also read how to create non-commodity content that AI cannot easily replicate for practical content-development strategies.
How Publishers Should Monitor the Situation
Gemini 3.8 Flash citation behavior is still a developing story. Publishers should monitor the trend instead of making major decisions from a small number of examples.
A structured tracking process will provide more useful information than occasional manual checks.
A 60–90 Day Monitoring Checklist
Start by selecting a fixed group of queries relevant to your website.
Include broad informational keywords, comparison searches, branded searches, and commercial-intent terms. Test the same queries regularly so you can compare changes over time.
Record whether AI Mode displays external links, source cards, citations, or brand mentions.
You should also monitor your analytics for changes in AI referral traffic. Compare those changes with the timing of major Google AI Mode or Gemini updates.
Pay particular attention to four signals:
- Whether Google publicly explains the reduction in citations.
- Whether future Gemini updates restore links to informational answers.
- Whether free-tier AI Mode users see the same citation behavior.
- Whether your own AI-referred sessions increase, decline, or remain stable.
A broader free-tier rollout would matter much more than the current limited exposure. It would place citation-free answers in front of a substantially larger search audience.
You can monitor brand presence with one of the AI visibility tools covered in our 2026 guide for every budget.
Manual testing also remains useful. Search the same category terms each month and keep screenshots of the responses. That creates a simple historical record of how Google’s citation behavior changes.
Frequently Asked Questions
What Is Gemini 3.8 Flash and Why Does It Matter for SEO?
Gemini 3.8 Flash is a Google AI model being used in AI Mode for certain users covered by this report.
The SEO concern comes from its citation behavior. Early examples show detailed answers to informational queries without visible external links or citations.
If Google expands that behavior to a much larger audience, publishers could receive fewer referral opportunities from informational searches.
Is Google AI Mode’s Citation-Free Behavior a Bug or a Feature?
Google had not publicly clarified the issue at the time covered by this report.
The missing citations could result from a temporary deployment problem, interface test, model behavior, or deliberate product direction.
Publishers should watch future updates before assuming the current implementation will become permanent.
Does Google-Extended Block Content From AI Mode?
The interaction between Google-Extended and every part of AI Mode is not fully transparent to publishers.
Google provides Google-Extended as a control for certain generative AI uses. However, publishers should verify which Google products and functions each directive governs before treating it as a universal AI Mode opt-out.
How Can I Track Whether AI Mode Is Affecting My Traffic?
Separate AI-referred visits from your other traffic channels in analytics.
Use GA4 reporting, referral data, and AI visibility monitoring tools. Compare monthly trends rather than relying on isolated daily changes.
You can also manually test your important queries in AI Mode and document whether your brand receives a mention, citation, or clickable link.
What Type of Content Is Most Protected From Zero-Citation AI Answers?
No content format can guarantee a citation.
However, original research, proprietary datasets, expert interviews, primary reporting, unique methodologies, and exclusive case studies create stronger differentiation than generic educational content.
AI can recreate a standard explanation from many available sources. Unique information gives the system fewer alternatives when it needs specific evidence.
Should I Opt Out of Google AI Mode to Protect My Traffic?
Do not make that decision from industry headlines alone.
Start with your own data. Determine how much visibility and referral traffic Google’s AI products currently provide to your site.
Then evaluate the benefits of appearing in AI-generated answers against the risks of reduced clicks and content reuse.
Google’s controls also differ across products and use cases, so publishers should verify what each available mechanism actually covers.
Conclusion
Google AI Mode running on Gemini 3.8 Flash is showing a notable citation pattern. Several top-of-funnel informational searches generate detailed answers without visible outbound links or source citations.
For publishers, the issue goes beyond a normal decline in organic CTR. A citation-free response can satisfy the user’s informational need while removing the direct path from the AI answer to the original publisher.
It is still too early to treat this behavior as permanent.
Publishers should measure their own AI referral traffic, track citation changes, and monitor Google’s future updates. At the same time, they should strengthen the parts of their strategy that do not depend on a single Google click.
Brand authority matters more when links disappear. Owned audiences reduce dependence on search. Bottom-of-funnel content can capture users with stronger action intent.
Most importantly, publishers should create information that cannot be reproduced from hundreds of interchangeable pages.
Original data, primary reporting, proprietary research, expert insight, and unique methodologies give your content a clearer reason to exist in an AI-driven search environment.
If the citation-free behavior expands to free-tier users, publishers may need to make larger strategic changes. Until then, consistent measurement provides a better foundation than panic-driven decisions.
To compare what you are seeing with other publishers, join Scale Xpert on Discord for SEO learning, publisher discussions, and genuine backlink exchange.
