December Core Update 2025: Specialized Brands Lead the Way

The December Core Update, which ran from December 11 to December 29, 2025, continued to influence search results into January 2026, showing clear winners and losers across industries. Specialized brands and niche publishers gained visibility, while broader, generalist sites, including large ecommerce platforms and news publishers, experienced ranking losses and volatility.

Who Gained Visibility

Brands and websites with category-specific expertise performed the best. These sites directly matched user queries, especially for “best of” and mid-funnel product searches.

• Ecommerce: Specialized retailers outperformed broad platforms. Columbia, The North Face, and Fragrance Market gained rankings for searches such as “winter boots women” and “men’s cologne,” while general retailers like Macy’s saw declines.
  
• SaaS Platforms: Niche software providers gained traction. Freshbooks and Xero, with dedicated landing pages for queries like “Accounting Software for Small Business,” rose in rankings. Non-specialized platforms such as Zapier and Adobe dropped.
  
• Brands: Companies with product authority, including Nintendo and Epic Games, increased visibility for gaming-related queries, overtaking general publications like Games Radar.

Who Lost Visibility

January 2026 data shows significant declines for news publishers and broad informational sites.

• News Publishers: India-based sites such as Hindustan Times, India Times, and Indian Express lost visibility in U.S. search results. Many experienced steep drops in Google Discover and Top Stories. Global news outlets were affected as well.
  
• Generalist Ecommerce and SaaS Platforms: Sites without category-specific authority continued to lose rankings for mid-funnel product and service searches.

Key Consequences for January 2026

• Traffic Shifts: Many generalist sites experienced traffic decreases, while specialized brands and retailers captured market share.
  
• Commercial Intent: Google appears to have reclassified certain “best of” queries as commercial, favoring sites with clear product or service authority.
  
• Planning Challenges for Publishers: Discover volatility in January 2026 created risks for news publishers without a niche focus. Traffic fluctuations were swift and impactful.

Lessons for Businesses and Marketers

• Focus on specialization and demonstrate expertise in narrow topic areas.
  
• Build dedicated landing pages and resource guides to strengthen authority.
  
• Monitor Discover and SERP trends closely to adapt to algorithm changes quickly.

The December Core Update continued to affect January 2026, emphasizing that niche authority, commercial relevance, and clear alignment with user intent are crucial for ranking success across all types of content.

---

Reddit Introduces Max Campaigns: Smarter Automation for Advertisers

Reddit has launched Max campaigns, a new automated campaign type now in beta for traffic and conversion objectives. This update aims to simplify campaign setup, improve performance, and provide advertisers with deeper audience insights, reflecting Reddit’s growing advertiser momentum and rising conversion activity.

Max campaigns automate key decisions including audience targeting, creative selection, placements, and budget allocation while staying within advertiser-defined guardrails. Powered by Reddit Community Intelligence, which analyzes over 23 billion posts and comments, Max campaigns predict the value of each ad impression in real time. This allows dynamic delivery adjustments based on performance, reducing the need for manual intervention. Optional creative automation tools further enhance campaigns by generating trending headlines, adapting images into Reddit-friendly formats, and soon offering AI-powered video cropping for repurposing content from other platforms.

Early tests with over 600 advertisers showed significant results. Campaigns achieved 17 percent lower cost per acquisition and 27 percent more conversions. Brooks Running, for example, reported a 37 percent drop in cost per click and 27 percent more clicks over 21 days without manual changes.

Unlike Google Performance Max or Meta Advantage Plus, Reddit combines automation with audience context, helping advertisers understand who engages with campaigns, what resonates, and how conversations shape results. Max campaigns are now available in beta with wider rollout expected in the coming months.

---

How Google’s Recommender Systems Now Understand User Intent

Google has introduced a new approach to recommender systems that helps them understand users’ semantic intent, meaning what people truly want when interacting with content. Traditional systems rely on clicks, ratings, or purchases, which are considered “primitive feedback” and often fail to capture subjective preferences, such as what a user finds funny, cute, or engaging. By leveraging Concept Activation Vectors (CAVs), Google’s system translates these subjective expressions into mathematical representations, allowing AI models to detect soft attributes, or nuanced personal preferences, without retraining the underlying model. This approach makes recommendations more personalized and context-aware, improving how platforms like YouTube, Google Discover, and Google News suggest content.

Benefits of this approach include:

• Detecting and understanding soft attributes like humor, tone, or style
  
• Distinguishing between objective attributes (e.g., genre, director) and subjective preferences
  
• Supporting interactive recommendations such as critiquing-based feedback
  
• Requiring minimal labeled data while allowing new attributes to be added easily
  
• Bridging the gap between human language and machine understanding, making recommendations more intuitive

Early tests show that this system can capture subtle, user-specific preferences, helping AI models provide more relevant and engaging content. While still in research stages, it has the potential to transform personalized content discovery and enhance overall user engagement.

---

Google Ads Deploys AI Model ALF to Detect Fraudulent Accounts

Google Ads has quietly deployed a powerful new AI model called ALF (Advertiser Large Foundation Model) to detect policy violations and fraudulent advertisers more effectively. Announced in a research paper dated December 31, 2025, ALF is a multimodal large foundation model that analyzes text, images, video, account history, billing details, and performance metrics together to understand advertiser behavior holistically.

Traditional systems often struggled with detecting malicious activity because individual signals, such as a declined payment or a newly created account, may seem harmless in isolation. ALF addresses this by combining multiple factors and comparing advertisers in large batches, allowing it to spot unusual behavior patterns using a technique called Inter-Sample Attention. This method enables the model to detect outliers that deviate from normal ecosystem activity.

• Handling heterogeneous and high-dimensional data from structured and unstructured sources
  
• Identifying malicious content hidden among thousands of creative assets
  
• Generating trustworthy confidence scores without excessive false positives
  
• Operating with strict privacy safeguards, stripping PII before processing
  
• Achieving over 40 percentage points improvement in recall and 99.8% precision on critical policies

While ALF’s model size increases latency slightly, it remains well within acceptable limits for production and serves millions of requests daily. Currently deployed within Google Ads Safety, ALF significantly improves fraud detection and policy enforcement, making the platform safer for advertisers and users. Future developments may expand its use for audience modeling, creative optimization, and evolving threat detection.

---

Google Reduces AI Overviews When Users Don’t Interact

Google’s AI Overviews appear selectively based on user engagement. Robby Stein, VP of Product at Google Search, explained that the system learns where Overviews are useful and reduces them for queries where users rarely interact. For example, searches for an athlete’s name focus on bios, photos, or social links, so AI Overviews often don’t appear.

The system also performs “under the hood” query expansion to surface relevant content, which can explain why citations may include pages that don’t exactly match the search terms. Overviews adapt to query type, integrating with images, products, or other relevant content.

AI Mode builds on this for more complex, conversational searches. Users can ask follow-up questions, making queries longer and more specific. Some personalization exists, like prioritizing video results for users who prefer them, but Google keeps the overall experience consistent.

This engagement-driven approach explains fluctuations in AI Overview presence, highlighting Google’s focus on usefulness and relevance rather than uniform display across searches.

---

Google Launches AI Mode Checkout and Branded Business Chat

Google has rolled out checkout in AI Mode and a new Business Agent feature, enabling shoppers to buy products and chat with brands directly from Search results. Eligible U.S. retailers remain the seller of record, while checkout happens on Google surfaces for a seamless experience.

The Universal Commerce Protocol (UCP) powers AI Mode checkout. Built with partners like Shopify, Etsy, Wayfair, Target, and Walmart, and supported by companies including Visa, Mastercard, and PayPal, UCP allows shoppers to pay via Google Pay or Google Wallet. Global expansion is planned.

Business Agent acts as a virtual sales associate, letting users ask questions and receive product guidance in the brand’s voice. Early adopters include Lowe’s, Michael’s, Poshmark, and Reebok, with future updates enabling in-chat purchases and personalized offers.

Google is also piloting Direct Offers, allowing retailers to show exclusive discounts to high-intent buyers. These features let users research, compare, and buy without leaving Google, giving participating retailers access to motivated shoppers while reshaping traditional traffic and sales patterns.

---

Google Downplays Chunking, But AI Search Results Remain Messy

Google downplays the need for “chunking” content for AI systems, emphasizing that well-structured web pages for human readers remain the best approach. On the Search Off The Record podcast, Danny Sullivan and John Mueller clarified that Google’s AI can access properly formatted content without publishers artificially breaking it into sections. The key takeaway is that writing for humans, not machines, remains the most effective SEO strategy, even in LLM-powered search.

Despite this guidance, AI search still produces low-quality SERPs, often promoting outdated, non-authoritative, or irrelevant content. For example, searches in Google AI Mode may surface abandoned blogs, LinkedIn posts, or retailer sites instead of expert sources like GQ or The New York Times. High-quality, authoritative content is increasingly hidden under the “More” tab or the News section, reducing traffic to publishers who provide real expertise.

The underlying issue is query fan-out, where Google ranks multiple pages for a single search query. This can dilute traffic for legitimate sites while amplifying low-quality results. Algorithmic debates like GEO or AEO don’t fully address the core problem: AI search often fails to surface trusted, expert content by default.

For SEOs and publishers, the focus should remain on creating human-first, authoritative content, while advocating for better AI search curation. Until Google addresses this properly, users and content creators will continue encountering subpar results instead of expert-driven answers.

---

AIOSEO WordPress Plugin Security Flaw Leaks AI Token

A critical vulnerability in the All in One SEO (AIOSEO) WordPress plugin has affected over 3 million websites, adding to six security issues disclosed in 2025. The flaw allowed low-privilege users, such as Contributors, to access a site’s global AI access token, potentially enabling unauthorized use of the plugin’s AI features, including generating content and consuming AI credits.

The issue stemmed from a missing capability check on the REST API endpoint /aioseo/v1/ai/credits. This endpoint, which reports AI usage and remaining credits, failed to verify whether the requesting user had permission. As a result, Contributors could retrieve the token meant only for administrators. Wordfence described this as a “site-wide credential leak,” which could be exploited to generate AI content or deplete service quotas, creating a potential denial-of-service scenario.

This vulnerability continues a pattern of permission-related flaws in AIOSEO, which had six disclosures in 2025 alone, including SQL injection, information disclosure, and unauthorized media deletion. By comparison, Yoast had zero, RankMath four, and Squirrly three vulnerabilities in the same year.

The vulnerability affects all versions up to AIOSEO 4.9.2 and has been fixed in version 4.9.3, which hardened API routes to prevent AI token exposure. Site owners are urged to update immediately, especially those with multiple external contributors, to secure their AI functionality.

---

OpenAI Starts Testing Ads in ChatGPT for Users in the U.S.

OpenAI is rolling out a limited test of advertising inside ChatGPT for Free and Go users in the United States. This marks the first time ads will appear within the ChatGPT experience, coinciding with the U.S. launch of ChatGPT Go, a low-cost subscription tier priced at $8 per month.

Ads will appear only at the bottom of relevant responses and will be clearly labeled and visually separated from ChatGPT’s answers. Users under 18, or queries related to sensitive topics like health, politics, and mental health, will not see ads. OpenAI emphasizes that ads will not influence responses, and user conversations will not be shared with advertisers. Users can also disable ad personalization entirely.

The new ad test is powered by carefully defined guardrails. OpenAI aims to maintain trust by ensuring answers remain unbiased and useful, rather than optimizing for engagement or revenue. Early ad formats include product recommendations under recipe suggestions and sponsored lodging listings for travel queries, all appearing below the response.

While this is not yet a full advertising platform, the test provides insights into how AI-first interfaces could introduce monetization without disrupting user trust. Paid tiers, including ChatGPT Pro, Business, and Enterprise, remain ad-free. OpenAI plans to adjust ad placement and formats based on user feedback as the U.S. test progresses, signaling a cautious approach to integrating ads into conversational AI.

---

More Sites Say No to LLM Crawlers – What It Means for Your AI Visibility

Recent analysis reveals a paradox: while companies are increasingly blocking LLM training bots, AI assistant crawlers are reaching more sites than ever.

Key Findings:

• AI assistant crawlers expanding: Coverage of websites by AI assistants grew significantly.
  
• LLM training bots restricted: Access for AI training bots dropped sharply over a few months.

Why It Matters:
LLM training bots build “parametric knowledge” — the long-term memory AI uses to answer questions. Blocking them can:

• Reduce first-party representation in AI answers.
  
• Force reliance on third-party data, misrepresenting your brand.
  
• Limit visibility of products, pricing, and messaging in AI responses.

Risks of Blocking LLMs:

• Users get answers without visiting your site → lower traffic and reduced brand control.
  
• Marketing attribution becomes harder as AI summaries replace site visits.
  
• Your content may be excluded from AI recommendations.

Strategic Approach:
Completely blocking LLMs may protect IP but can limit visibility and influence. Selective, controlled access ensures AI models represent your brand accurately while safeguarding sensitive content.

---

Google Introduces Personal Intelligence for AI Mode

Google has rolled out Personal Intelligence, connecting Gmail and Google Photos to AI Mode for personalized responses. The feature is available as a Labs experiment for AI Pro and AI Ultra subscribers in the U.S., limited to personal accounts.

Key Features

• AI Mode can reference emails and photos to tailor responses.
  
  • Travel suggestions can consider bookings and past travel photos.
    
  • Shopping recommendations can factor in preferences, upcoming trips, and weather.
    
• Reduces the need for users to provide repeated context.

How to Enable

1. Open Google Search → tap profile
   
2. Click Search personalization
   
3. Select Connected Content Apps
   
4. Connect Gmail and Google Photos

Privacy & Controls

• Opt-in feature, can be turned off anytime.
  
• AI doesn’t train on Gmail or Photos; only prompts and responses are used for improvements.
  
• Users can correct mistakes via feedback.

Why It Matters

The feature delivers personalized answers, enhancing convenience for users. For publishers, some queries may resolve within AI Mode, potentially reducing clicks to external sites.

Personal Intelligence will roll out gradually. Expansion to free or Workspace accounts could broaden reach, and publishers should monitor its impact on content visibility and traffic.

---

Google AI Overviews Now Run on Gemini 3

Google has upgraded AI Overviews to use Gemini 3 as the default model, bringing advanced reasoning capabilities previously available in AI Mode directly to search results. The update also allows users to seamlessly continue from an AI Overview into AI Mode, carrying over context for follow-up questions.

What’s New: Gemini 3 Powers AI Overviews
Gemini 3, launched in November, now extends its reasoning abilities from AI Mode into AI Overviews. This upgrade aims to provide more accurate, context-aware responses to complex queries. The feature currently reaches over 1 billion users globally.

Seamless Transition to AI Mode
Users can now start a follow-up question directly from an AI Overview without losing context, creating a smooth, conversational search experience. This approach helps users explore topics deeply without restarting their queries.

Using the same model across AI Mode and AI Overviews ensures consistency in reasoning, citation patterns, and response quality. For content creators and publishers, this update may influence which pages are cited and how information is structured in AI-powered responses.

The rollout is underway, with availability varying by region. Google has indicated plans for automatic model selection, routing complex queries to Gemini 3 while using faster models for simpler tasks, which could further shape AI Overview behavior in the future.

---

Chrome Adds 3 New AI Features, Including Nano Banana

Google has upgraded Chrome with three new AI features, integrating Gemini capabilities on Windows, MacOS, and Chromebook Plus. The update introduces an AI side panel, agentic AI Auto Browse, and Nano Banana in-browser image editing. The AI side panel allows users to run a Gemini session without switching tabs, making it easier to multitask, compare options across multiple sites, summarize reviews, and manage busy schedules. Users must opt in and consent to sharing browser data and URLs to enable this feature.

Nano Banana enables users to edit images directly in the browser window without downloading or uploading, streamlining image updates instantly. Autobrowse, available to AI Pro and Ultra subscribers, allows an AI agent to act on the user’s behalf, performing tasks like researching flights, hotels, and products, comparing costs, adding items to carts with discount codes, and even logging into accounts if granted permission. Its multimodal functionality can identify items in images and find purchase options automatically.

These features complement Chrome’s existing AI integrations with apps like Calendar, Gmail, Google Shopping, Flights, Maps, and YouTube, part of Google’s Personal Intelligence initiative, which aims to create a more personalized AI assistant experience. Google’s research also shows that in-browser and on-device AI can detect user intent to provide proactive, context-aware responses, pointing toward a future of smoother, more tailored AI interactions in Chrome.

---

ChatGPT Prism: AI-Powered Research Made Smarter

In January 2026, OpenAI launched ChatGPT Prism, a new AI tool designed to transform research and content workflows. Unlike traditional ChatGPT, Prism goes beyond answering questions, offering structured summaries, data-driven insights, and actionable recommendations.

Key Features

• Multi-document synthesis: Prism can analyze multiple sources at once and highlight trends or conflicting information.
  
• Semantic understanding: It interprets context and user intent to provide highly relevant insights.
  
• Customizable workspace: Users can set focus areas, output formats, and priorities, making it suitable for research, marketing, content creation, and data analysis.

Who Benefits

• Researchers: Quickly compile reports from various sources without manual aggregation.
  
• Content creators and marketers: Generate outlines, topic clusters, and data-backed content recommendations in minutes.
  
• Business analysts: Extract actionable insights from large datasets or reports efficiently.

Prism reduces time spent on manual work, allowing professionals to focus on strategy, analysis, and decision-making.

Implications for SEO and Content

As AI-generated search grows, Prism highlights the importance of:

• Authoritative content: High-quality, structured information is more likely to be used in AI summaries.
  
• Depth and relevance: Content optimized for semantic intent performs better in AI-driven results.
  
• Faster workflows: Teams can research, plan, and create content more efficiently, gaining a competitive edge.

ChatGPT Prism represents a shift from conversational AI to strategic AI, integrating deeply into workflows. Businesses and marketers that adopt it early can boost productivity, create high-quality content, and better meet user intent, signaling a new era of AI-assisted research.

---

FAQs

---

Conclusion

January 2026 has been a landmark month for AI, search, and digital marketing. From Google’s Gemini 3-powered AI Overviews and Personal Intelligence to ChatGPT Prism’s advanced research capabilities and Chrome’s new AI tools, personalization, reasoning, and automation are reshaping user experiences. OpenAI’s cautious ad testing and the rise of AI assistant crawlers highlight the growing importance of strategic AI visibility for businesses. Meanwhile, SEO trends emphasize specialization, authority, and alignment with user intent as key factors for ranking success. Staying updated with these developments is essential for marketers, publishers, and brands seeking to maintain relevance and maximize reach in an AI-driven digital landscape.

As AI continues to integrate deeper into search, browsing, and content creation, businesses must rethink how they engage audiences. Visibility is no longer just about clicks and impressions, it’s about ensuring AI systems understand and represent your brand accurately. Brands that adopt AI strategically, optimize for semantic intent, and maintain authoritative, high-quality content will gain a competitive edge, while those ignoring AI-driven trends risk losing influence in both search and AI-powered experiences.