Catch up on select AI news and developments from the past week or so:
Google launches Gemini 3 and bakes it into search from Day One. Google has unveiled Gemini 3 and immediately wired it into its core money-makers, including search and paid AI Mode for complex queries. The model claims top scores on industry leaderboards but, unlike earlier releases, is already powering revenue-generating consumer and enterprise products. Google also introduced Gemini Agent for multistep tasks like inbox triage and travel booking, plus Antigravity, a new AI-driven software development platform. A revamped Gemini app now returns rich, website-like results, further disintermediating publishers by answering in a self-contained interface. Importance for marketers: Gemini 3's default presence in search and AI Mode raises the stakes for GEO/AEO, rich answers, and AI-era SEO, with more traffic staying offsite in AI-native experiences.
Trump weighs executive order to override state AI rules. President Donald Trump is considering an executive order to undercut state-level AI laws using federal lawsuits and funding pressure. A proposed AI Litigation Task Force under Attorney General Pam Bondi would challenge state regulations as burdensome or unconstitutional, while the Commerce Department could tie broadband funding to AI policy. The draft criticizes California's disclosure law and Colorado's algorithmic discrimination rules as heavy-handed. The push follows earlier congressional efforts to block state AI regulations via the defense budget, which drew bipartisan resistance. Importance for marketers: National pre-emption could reshape AI compliance, narrowing state-level guardrails on deepfakes, targeting, and automated decision-making, and shifting risk management toward evolving federal rules instead.
OpenAI faces privilege challenge that could boost copyright damages. Authors and publishers suing OpenAI over the use of pirated books have secured internal messages about a deleted Library Genesis dataset and now seek attorney communications tied to that decision. Plaintiffs argue OpenAI waived privilege by previously giving a benign explanation for the deletion and are invoking the crime-fraud exception as well. If the court forces disclosure and finds willful infringement or spoliation, OpenAI could face enhanced damages of up to $150,000 per work, sanctions, or adverse jury instructions. Importance for marketers: The case could set precedents for training data transparency and liability, affecting how marketing teams evaluate AI vendors, copyright risk, and IP-safe content workflows.
TikTok adds controls and invisible watermarks for AI-generated content. TikTok is introducing a new AI-generated content control inside its Manage Topics tool, letting users dial AI content up or down in their For You feed. The company is also testing invisible watermarking on AI outputs, complementing C2PA metadata so detection persists even after edits or reuploads. Watermarks will apply to AI made with TikTok tools and to uploads that already carry Content Credentials. TikTok is also launching a $2 million AI literacy fund with partners like Girls Who Code. Importance for marketers: Expect more pressure to clearly label synthetic media, plus more nuanced reach dynamics as users tune AI content exposure, influencing creative choices and disclosure practices.
Adobe buys Semrush to deepen AI-powered marketing analytics. Adobe is acquiring Semrush for $1.9 billion in cash, a 77.5% premium, to bolster its AI-driven marketing and analytics stack. Semrush's platform helps brands optimize search, social, and digital advertising performance. Adobe plans to combine Semrush data with its creative and marketing tools so brands can see how they appear across traditional search and generative AI bots like ChatGPT and Gemini. Analysts call the price steep but strategically meaningful, even as questions linger about Creative Cloud's growth story. Importance for marketers: This move signals tighter integration between content creation, SEO, and AI-driven visibility intelligence, potentially making Adobe a more end-to-end hub for performance and brand measurement.
OpenAI and Target expand AI partnership with ChatGPT shopping app. Target is launching a new app inside ChatGPT that lets customers ask for ideas, browse curated recommendations, build multi-item baskets, and check out via Drive Up, Order Pickup, or shipping. The retailer already uses ChatGPT Enterprise for 18,000 employees and has built AI tools like Agent Assist, Store Companion, Shopping Assistant, Gift Finder, and JOY to streamline service, support, and vendor interactions. The new ChatGPT experience will later add features like Target Circle linking and same-day delivery. Importance for marketers: This showcases conversational commerce as a real retail channel, pushing brands to design AI-native shopping journeys, prompts, and merchandising strategies inside third-party assistants.
LinkedIn rolls out AI search to find the right people by description. LinkedIn is deploying an AI-powered people search that lets users describe whom they seek in natural language instead of entering names or strict filters. Queries like "Northwestern alumni who work in entertainment marketing" or "Who can help me understand the US work visa system?" are translated into ranked results based on connection strength and relevance. The feature builds on LinkedIn's AI-enhanced job search and is rolling out first to Premium users in the US, with broader access planned. Importance for marketers: Prospecting, influencer discovery, and partner outreach become more intent-driven, rewarding clear positioning and profile optimization aligned with AI-readable expertise and signals.
Microsoft introduces Agent 365 to monitor and control AI agents at work. Microsoft has announced Agent 365, a management layer for autonomous AI agents that automates office tasks. The tool gives IT teams visibility into agents across systems, lets them quarantine misbehaving ones, and governs access to resources, whether agents are built on Microsoft platforms or external tools like Salesforce. Launched at Microsoft Ignite, Agent 365 arrives alongside Work IQ, which lets companies build agents atop the same intelligence and business data powering Microsoft 365 Copilot. Importance for marketers: As agentic workflows spread across CRM, analytics, and content ops, governance tools like this will shape how marketing teams adopt automation safely and measure ROI.
Affirm CEO says agentic AI will reshape shopping and payments. Affirm co-founder and CEO Max Levchin argues that AI agents will soon read fine print, filter out predatory products, and guide consumers toward transparent financial options. Speaking at Reuters Momentum AI Finance, he said agentic AI will "weed out" fee structures designed to trap inattentive shoppers. The comments come as buy now, pay later spending hit $82.4 billion in 2024, up 9.9% year over year, and as retailers like Walmart plan AI "super agents" to power e-commerce and operations. Importance for marketers: If AI shoppers evaluate terms ruthlessly, brands will need clearer pricing, honest offers, and agent-friendly product data to stay in shortlists and avoid being screened out.
Prime Video adds AI-generated video recaps for select series. Amazon's Prime Video is rolling out AI-generated Video Recaps that summarize previous seasons or episodes with synchronized narration, dialogue, and music. Initially available in beta for shows like Fallout, Tom Clancy's Jack Ryan, and Upload, the feature builds on last year's X-Ray Recaps, which offered AI summaries while trying to avoid spoilers. Competitors are experimenting too: YouTube TV's Key Plays and Netflix's production tools show how generative AI is entering viewing and content creation. Importance for marketers: Viewers may come to expect fast, cinematic summaries and personalized recaps, opening new opportunities for promo slots, branded storytelling, and recap-style formats across streaming and owned channels.
Google launches Nano Banana Pro image model built on Gemini 3. Google is releasing Nano Banana Pro, officially called Gemini 3 Pro Image, as its new flagship image generation and editing model. Built on Gemini 3 Pro, it offers sharper text rendering, better multilingual typography, more flexible styles, and smarter controls for angles, focus, lighting, and color grading. It supports localized edits, multiple aspect ratios, and up to 4K resolution, making it useful for infographics and diagrams. The model is rolling out globally in the Gemini app, AI Mode, NotebookLM, Workspace, and developer tools. Importance for marketers: Higher-fidelity, controllable visuals and infographics make it easier to produce on-brand creative variations and data visuals fast, while evolving watermark policies keep authenticity checks in play.
Meta unveils SAM 3 and SAM 3D for precise editing and 3D reconstruction. Meta is expanding its Segment Anything suite with SAM 3 for advanced image and video segmentation and SAM 3D for reconstructing objects and people in three dimensions. SAM 3 can detect and track objects from detailed text prompts, while SAM 3D generates 3D assets from 2D inputs using specialized models for objects and human bodies. Meta is testing the tech in its Edits video app, Vibes short-form platform, and Facebook Marketplace's View in Room feature. Parts of SAM 3 are fully open-sourced. Importance for marketers: These tools point to faster creation of product cutouts, AR try-ons, and 3D brand assets, enhancing shoppable experiences and visual personalization in ads and marketplaces.
NotebookLM adds Deep Research and support for more document types. Google's NotebookLM is gaining a Deep Research mode that plans and executes multi-step web research, then returns source-grounded reports users can drop straight into their notebooks. Users can choose Deep Research for full briefings or Fast Research for quicker passes. NotebookLM now also supports Google Sheets, Drive file URLs, PDFs from Drive, and Microsoft Word documents, enabling summaries and analysis across more formats. These features layer on existing Video and Audio Overviews and mobile apps. Importance for marketers: The upgrade strengthens AI-powered research, briefing, and content development workflows, making it easier to synthesize data, campaign inputs, and stakeholder docs into coherent presentations and narratives.
Google tests Scholar Labs AI tool to surface research papers. Scholar Labs is a new AI-powered research search tool from Google that analyzes full-text content, topics, and relationships to answer detailed scientific queries. Unlike classic Google Scholar, it does not let users filter results by citation count or journal impact factor, though those signals still inform ranking. Instead, it prioritizes semantic relevance across fields and recency. Scientists interviewed note that impact metrics are crude but still useful, raising questions about how researchers will vet credibility in AI-ranked results. Importance for marketers: Thought leadership and research-based content strategies may evolve as AI surfaces different studies and sources, nudging teams to scrutinize methodology and quality, not just citation prestige.
Major labels license music catalogs to AI streaming startup Klay. Universal Music Group and Warner Music Group have struck licensing deals with Klay, an AI-first music service that blends traditional streaming with tools for users to remake songs in new styles. Klay is reportedly the first AI music platform to secure deals with all three major labels and has licensed thousands of tracks for training, while promising artists and labels control over how works are used. The move comes amid numerous copyright lawsuits against AI music firms, including a recent Warner settlement with Udio. Importance for marketers: Licensed AI music remixing could normalize brand-safe, customizable soundtracks, but it will also raise expectations for rights-cleared, adaptive audio in campaigns and experiences.
Anthropic details how it measures Claude's political neutrality. Anthropic has published how it aims to keep its Claude chatbot politically even-handed, following President Trump's ban on so-called "woke AI" for government procurement. Claude's system prompt instructs it to avoid unsolicited political opinions, maintain factual accuracy, and represent multiple perspectives. Reinforcement learning rewards traits such as responses that do not clearly signal conservative or liberal leanings. Anthropic also built an open-source tool to score political even-handedness, where Claude models outperformed rivals like GPT-5 and Meta's Llama 4. Importance for marketers: As political and issue-based content faces tighter scrutiny, understanding how models try to manage bias will matter for campaigns, advocacy work, and brand safety in sensitive discussions.
Databricks co-founder warns US needs open source to stay ahead of China in AI. Andy Konwinski, Databricks co-founder and now head of the Laude Institute, argues that the US is ceding AI thought leadership to China as more groundbreaking ideas and papers emerge from Chinese labs. He says proprietary strategies at big US AI firms are draining universities of talent and limiting open exchange. Konwinski cites the original Transformer paper as proof that open research drives breakthroughs and notes that Chinese models like DeepSeek and Alibaba's Qwen are encouraged to be open. Importance for marketers: The balance between open and closed AI ecosystems will influence which tools, models, and communities shape the future of marketing innovation and experimentation.
OpenAI launches GPT-5.1-Codex-Max for long-running, agentic coding tasks. OpenAI has introduced GPT-5.1-Codex-Max, a new high-context coding model optimized for long-horizon software development. Using a compaction mechanism, it can retain key context while dropping irrelevant details, enabling work across millions of tokens and even 24-hour internal tasks like multi-step refactors and debugging. Benchmarks show notable gains over GPT-5.1-Codex on SWE-Bench, Terminal Bench 2.0, and other software engineering tests. The model is live in OpenAI's Codex CLI and internal tools, with public APIs promised later. Importance for marketers: More capable coding agents will accelerate shipping AI-powered marketing tools, integrations, and experiments, even for smaller teams that rely heavily on AI for engineering capacity.
xAI's Grok 4.1 boosts personality, empathy, and factual accuracy. xAI has fully released Grok 4.1 across grok.com, X, and mobile apps after a two-week silent rollout. The model focuses on real-world usability, improving creativity, emotional intelligence, and conversational style while maintaining strong reasoning. Large-scale reinforcement learning tuned traits like helpfulness and alignment, leading to a 64.78% win rate over the prior model in blind user evaluations. Grok 4.1 also shows significant reductions in hallucinations on information-seeking prompts and tops several public capability and emotional intelligence benchmarks. Importance for marketers: As personality-rich, less hallucinatory models spread, brands gain more options for on-voice assistants, social engagement bots, and emotionally sensitive customer interactions.
Alibaba rebrands Tongyi into Qwen app to chase consumer AI users. Alibaba has launched the Qwen App, a revamped consumer chatbot based on its latest large language model, in a renewed push into consumer AI. Available as a free mobile and web app in China, Qwen can generate research reports and polished slide decks from a single command and is marketed as a powerful personal assistant. The move represents a strategic shift from Alibaba's enterprise-focused cloud AI, as it seeks to catch rivals like ByteDance's Doubao, DeepSeek, and Tencent's offerings in a highly competitive domestic market. Importance for marketers: Qwen's capabilities highlight the rising bar for AI assistants in China, affecting how global brands plan localized experiences and compete for attention in AI-first consumer channels.
Adobe and Qualcomm back Humain's Arabic-focused generative AI push. Adobe and Qualcomm are partnering with Humain, a Saudi-backed AI company, to build generative AI tailored to Arabic and Middle Eastern contexts. Adobe will integrate Humain's Allam large language model into its creative and marketing apps, while Humain will use Adobe Firefly Foundry to develop region-specific models. Qualcomm's new AI200 and AI250 chips will power Humain's data centers, with a joint R&D center opening in Riyadh and plans to deploy 200 megawatts of AI chips next year. Importance for marketers: Expect more culturally fluent, Arabic-native creative tools and campaigns, raising expectations for regional personalization, local-language content, and AI-powered storytelling across the Middle East.
Putin calls for national AI task force to secure Russian sovereignty. President Vladimir Putin is establishing a national task force to coordinate Russia's generative AI efforts, arguing that relying on foreign large language models is unacceptable for a sovereign nation. He wants Russian-made models for national security and intelligence, expanded data centers supported by small nuclear plants, and AI's contribution to GDP to exceed 11 trillion rubles by 2030. Despite Western sanctions limiting hardware imports, Russia's Sberbank and Yandex have launched major LLMs like Gigachat and Yandex GPT. Importance for marketers: Growing AI nationalism and local-model mandates may fragment the global AI landscape, pushing international brands to navigate country-specific platforms, standards, and content rules.
You can find the previous issue of AI Update here.
Editor's note: GPT-5.1 was used to help compile this issue of AI Update.
