Catch up on select AI news and developments from the past week or so:

New York law requires advertisers to disclose AI-generated people. New York Gov. Kathy Hochul signed a law requiring advertisers to clearly disclose when an ad includes AI-generated people or synthetic avatars that do not correspond to a real human. A separate bill also requires consent from heirs or executors to use a deceased person's name or likeness commercially. The disclosure requirement must be "conspicuous," with exceptions for ads promoting expressive works (movies, TV shows, video games) where the avatar use aligns with the work itself. The change reflects rising scrutiny of AI use in advertising. Importance for marketers: This creates a concrete compliance obligation for AI creative. Teams need disclosure rules, review checklists, and vendor guidance now, especially if using synthetic spokespeople, influencer replicas, or avatar-led performance ads.

Google makes Gemini 3 Flash the default model across Search AI Mode and Gemini. Google released Gemini 3 Flash and is pushing it broadly: It becomes the default model in the Gemini app for everyday tasks and the default behind AI Mode in Search, exposing global users to the model at scale. Google positions Flash as faster, cheaper to run, and equipped with Gemini 3 Pro reasoning. It also highlights multimodal capabilities (video, images, audio, text) that can be turned into content. The release lands amid intensified competition with OpenAI and growing pressure to maintain accuracy at scale. Importance for marketers: AI answers inside Search further shift discovery away from blue links. Marketers need stronger authority signals, clearer structure, and content designed to be cited or summarized accurately in AI-driven results.

ChatGPT Images expands OpenAI's competition with Google and Adobe. OpenAI launched ChatGPT Images for generating and editing images, with faster generation and more precise edits that preserve details like lighting, composition, and consistent appearance across iterations. The feature appears in a new ChatGPT sidebar for all users and is available in the API as GPT Image 1.5. The release responds to enthusiasm for Google's Nano Banana image model, which has been praised for studio-quality outputs and strong text rendering and has already been integrated into Adobe Firefly and Photoshop. Importance for marketers: Image creation and editing are becoming default capabilities inside mainstream AI tools. Creative velocity will rise, but so will sameness and IP risk, making brand controls, review processes, and creative differentiation more important.

AI content experiments trigger backlash across publishers and brands. Brands and publishers are facing backlash as AI-driven content experiments produce factual errors and quality failures. Examples include Amazon pulling an AI-generated Prime Video recap after mistakes, and McDonald's Netherlands removing an AI clip-based Christmas ad after online criticism. Newsrooms have repeatedly stumbled since 2022, from inaccurate AI articles to fabricated quotes and even printed lists featuring nonexistent book titles attributed to real authors. Successful AI integrations require rigorous testing, editorial review, transparency, and staff buy-in as lawmakers and unions increase scrutiny. Importance for marketers: This is a warning: Rushed AI content can damage trust quickly. Marketers should formalize human review, provenance standards, and disclosure policies before scaling AI-generated copy, audio, or video.

AI search tools raise the bar for SEO rather than replacing it. AI-powered search and generative answer tools are changing discovery, but the piece argues they do not eliminate SEO. Core factors like authority, clarity, site structure, depth, and trust still influence visibility, including in AI-driven answer engines. The shift broadens SEO into two arenas: classic rankings and AI-generated responses. While AI can speed analysis and draft work, human specialists remain essential for strategy, intent interpretation, credibility, and creating content that resonates and aligns with brand goals. The article emphasizes E-E-A-T-like expectations and trust as central. Importance for marketers: Treat SEO as a strategic, trust-led discipline across search and answer engines. Invest in content quality, structure, and credibility building, not only tooling or automation.

Creative Commons warns pay-to-crawl could become a new web chokepoint. Creative Commons outlines how pay-to-crawl systems might compensate sites when machines access content, but warns the approach can become extractive, DRM-like, or power-concentrating if implemented poorly. CC supports nuanced controls that distinguish between different machine users (commercial AI, nonprofits, researchers) and different purposes (training, indexing, inference). It argues these systems should preserve public-interest access, avoid surveillance, allow throttling (not only blocking), and use open, interoperable standards rather than proprietary gatekeepers. The goal is reciprocity without undermining the open web. Importance for marketers: If pay-to-crawl expands, content distribution, measurement, and discoverability could change fast. Marketing leaders should track how their sites handle crawlers and how content licensing might affect AI visibility.

Adobe Firefly adds more control to AI video creation and editing. Adobe updated Firefly with new generative video features and partner model access aimed at reducing "prompt roulette" and increasing practical control. Prompt to Edit supports precise natural-language edits to existing clips, while new camera motion controls allow users to guide movement via a reference video. Firefly also adds model choices such as Topaz Astra for upscaling and FLUX.2 for image generation and editing with strong text rendering and reference image support. Adobe is offering unlimited generations for subscribers through January 15, 2026. Importance for marketers: More controllable AI video makes scaled creative testing more realistic. Teams can iterate faster, but should still enforce brand standards, legal review, and asset governance for multi-model workflows.

Zoom AI Companion 3.0 expands from meeting summaries into agentic workflows. Zoom launched AI Companion 3.0, pushing beyond meeting summaries into browser access, agentic workflows, low-code automation, and federated orchestration across multiple AI models. The assistant can retrieve information across meetings, notes, and connected third-party apps, generate documents via "Help Me Write," and support a drag-and-drop workflow designer for agent-driven processes. Zoom describes a federated approach that starts with small models and escalates to larger models when needed to balance cost and quality. Pricing is included with paid Zoom Workplace licenses, with additional tiers available. Importance for marketers: Marketing work is increasingly executed inside collaboration suites. Agentic automation in meeting and document workflows can speed briefs, recaps, and follow-ups, but needs governance around data access and output quality.

Microsoft Promptions replaces prompt trial-and-error with guided options. Microsoft introduced Promptions (prompt + options), an open-source UI framework meant to reduce the frustrating loop where prompts miss the mark and users keep rephrasing. Promptions uses conversation context to generate clickable controls such as response length, tone, focus areas, and format, helping users express intent without writing long, precise prompts. Tests found dynamic controls made it easier to specify tasks and reduced prompt fatigue, though users sometimes found the controls opaque and had to learn how options affected outputs. The design is stateless, reducing data governance complexity. Importance for marketers: Standardized UI controls can improve consistency across teams using AI for writing, analysis, and enablement. That reduces dependence on individual prompt skill and supports repeatable best-practice workflows.

Google Translate adds real-time headphone translations and Gemini improvements. Google is rolling out a beta feature that lets users hear real-time translations through headphones, while preserving each speaker's tone, emphasis, and cadence so conversations are easier to follow. The feature works with any headphones, supports 70+ languages, and is launching on Android in the US, Mexico, and India, with iOS and broader country support planned for 2026. Google is also adding Gemini-powered translation improvements for slang, idioms, and nuanced phrasing, plus expanded language learning tools and progress tracking across more countries and languages. Importance for marketers: Real-time translation reduces friction for global engagement and support. Brands should raise their localization standards and consider how multilingual experiences and content can convert in more markets.

Wearables pivot toward always-on AI as the core value proposition. The article argues 2025 marked a shift: wearables are increasingly positioned as vehicles for AI rather than primarily health and fitness devices. "Smart glasses" are being rebranded as "AI glasses," with Meta calling glasses the ideal form factor for AI due to on-face audio and camera context. Beyond glasses, always-listening pendants and pins are appearing to summarize meetings and conversations, generate to-do lists, or act as companions. Meanwhile, major platforms are embedding AI into watches, rings, and earbuds through assistants, coaching, and translation features. Importance for marketers: AI wearables expand context-aware touchpoints and change how people consume information. Marketing should prepare for ambient discovery and higher sensitivity around privacy, consent, and relevance.

Meta adds conversation boosting and context-based music to AI glasses. Meta is rolling out Conversation Focus to Ray-Ban Meta and Oakley Meta HSTN glasses for Early Access users, using directional microphones to amplify the voice of the person you are speaking with in noisy environments. Users can adjust the feature via swipes on the glasses or device settings. Meta is also adding Spotify integration so Meta AI can play music based on what users are looking at, framed as an environment-matching holiday feature. The piece notes Google has demoed similar context-based music experiences for prototype XR glasses. Importance for marketers: Glasses as an AI interface move brands closer to real-world, context-driven experiences. It is early, but it points to future ad and commerce formats that rely on situational relevance rather than clicks.

Companies are not yet seeing strong returns from AI investments. Reuters reports many businesses are struggling to realize meaningful ROI from generative AI, despite widespread adoption since ChatGPT's launch. Executive surveys cited include Forrester finding 15% saw profit margins improve due to AI and BCG finding only 5% saw widespread value. Firms cite problems such as model "sycophancy," inconsistent performance, failures on seemingly simple tasks, and difficulty working with long documents. Some companies are reverting to more human support for complex customer interactions. AI vendors are responding with more hands-on deployment support and business-focused offerings. Importance for marketers: This is a caution against assuming AI is an "easy button." Marketing leaders should set realistic expectations, focus on high-impact, low-lift workflows first, and measure outcomes like cycle time, pipeline support, and quality.

Nvidia releases Nemotron 3 Nano as open source model competition intensifies. Nvidia unveiled the third generation of its Nemotron large language models and released the smallest version, Nemotron 3 Nano, saying it is faster, cheaper to run, and better at long, multistep tasks than predecessors. Two larger versions are planned for the first half of 2026. The release comes as open source models from Chinese AI labs proliferate, with some companies disclosing use of models like Alibaba's Qwen. Nvidia emphasized dependability, releasing training data and tools so governments and enterprises can test security and customize models. Importance for marketers: Open source model options can enable more customized brand experiences and reduce platform dependency. It also increases the need for internal evaluation and governance to choose models that fit quality and security needs.

Google upgrades Gemini Deep Research and opens embedding via Interactions API. Google released a "reimagined" Gemini Deep Research agent based on Gemini 3 Pro. Beyond generating research reports, it now supports developers embedding research capabilities into their own apps through Google's new Interactions API, aimed at an agentic era where AI systems do more information seeking. Google says customers use the tool for due diligence and scientific research. It claims improved factuality and reduced hallucinations and introduced a new DeepSearchQA benchmark, while also testing against external benchmarks. The launch timing coincided with OpenAI's GPT-5.2 release. Importance for marketers: Research agents will increasingly shape how decisions are made and which sources get surfaced. Marketers should optimize for being cited, summarized, and trusted in agent-driven research workflows, not only search rankings.

Notion's AI products drive growth and an employee share sale at an $11B valuation. Notion enabled employees to sell shares at an $11 billion valuation, with investors including Sequoia, Index, and GIC reportedly participating. The company recently passed $600 million in annual recurring revenue and is described as cash flow positive, with roughly half of ARR attributed to AI products. Notion is expanding beyond a flexible productivity tool into an "everything app" for office work, potentially challenging dominant suite providers. AI features cited include enterprise search, a knowledge work agent that can perform extended tasks, and an AI note taker. Importance for marketers: Marketing operations are increasingly built inside productivity platforms. AI-native knowledge and workflow tools can improve planning, enablement, and execution speed, but teams should standardize how information and decisions are captured.

Trump administration launches Tech Force to bring Big Tech into government modernization. The White House announced an AI-focused "Tech Force" program to recruit technologists, including from major tech firms, for two-year stints to modernize federal government systems. Participating company partners listed include Adobe, AWS, IBM, Meta, Microsoft, Nvidia, Oracle, Palantir, and xAI. The program targets 1,000 recruits with salaries between $135,000 and $195,000, and aims to improve government capacity after workforce disruptions and cuts reduced existing expertise. Participants must follow ethics rules, and security clearance processes will apply. The initiative echoes earlier modernization efforts like US Digital Service. Importance for marketers: Government AI modernization can influence procurement, standards, and public-sector adoption patterns. Marketers selling into regulated or public markets should watch how these partnerships reshape requirements and credibility signals.

Trump executive order seeking to block state AI laws faces major hurdles. Reuters reports President Trump's executive order aimed at barring state AI laws will likely face political and legal opposition. The order directs agencies to sue and withhold funds from states with AI laws deemed problematic, including using the $42 billion BEAD broadband program as leverage. Legal experts question whether the administration has authority to tie broadband funding to AI policy, especially where states have already received approvals. Some Republicans and governors have criticized federal efforts to override state AI rules. Courts will likely assess dormant commerce clause arguments and whether states are treated unequally. Importance for marketers: AI regulation remains fragmented and unstable. Marketing teams using AI for personalization, content generation, or targeting should build flexible compliance and documentation practices that can adapt across states.

Disney pressures Google to remove AI videos using Disney characters and add safeguards. Variety reports Google removed dozens of AI-generated YouTube videos depicting Disney-owned characters after Disney sent a cease and desist letter. The videos included characters such as Mickey Mouse, Deadpool, and figures from Star Wars and The Simpsons, many created using Google's Veo AI video tool. Disney demanded removals, requested safeguards to prevent Google's AI tools from generating Disney characters, and asked Google to stop using Disney characters for training. The action came shortly before Disney announced a deal to license 200 characters to OpenAI for Sora users, underscoring a controlled-licensing approach. Importance for marketers: IP enforcement around AI-generated characters is tightening. Brands should avoid generating recognizable characters or styles without rights, and should review creative workflows for infringement and training-data concerns.

IBM releases CUGA agent that completes more than half its tasks in benchmarks. IBM researchers released CUGA (Configurable Generalist Agent), an open source agent designed to automate complex enterprise workflows through orchestration, API integration, and code generation. The Register notes benchmark results such as 61.7% success on WebArena and 48.2% on AppWorld, with commentary that these rates might be acceptable only in specific scenarios. The write-up highlights broader skepticism about agents, including warnings that many enterprise agent projects may be canceled for lack of business value. CUGA uses intent detection, a task ledger, delegation to specialized agents, and policy-aware execution. Importance for marketers: Agents are improving but remain inconsistent. Marketing should treat agentic automation as assistive, with strong guardrails and QA, especially for tasks that can create customer-facing errors or compliance risk.

Vanguard analysis suggests AI-exposed jobs saw higher wage and job growth. Vanguard analyzed Labor Department occupational data and found that jobs with high AI exposure saw stronger wage and job growth from Q2 2023 to Q2 2025 than roles with low exposure. Real wages rose 3.8% in high-exposure occupations versus 0.7% in others, while job growth increased 1.7% versus 0.8%. The analysis argues AI is augmenting work, enabling higher-value activity and, in some cases, increasing demand, with software engineering cited as an example where lower costs can expand use cases. Economists caution this is a snapshot, not a long-term forecast. Importance for marketers: This supports planning for productivity gains and role evolution rather than panic-driven cuts. Marketing leaders can justify training and tooling investments that expand capacity and speed experimentation.

Mozilla signals Firefox will become an AI browser with opt-out AI features. Mozilla's new CEO outlined a plan to evolve Firefox into a "modern AI browser," positioning AI as a path to diversify revenue as market share declines and the Google search deal becomes less secure. The piece describes an upcoming Firefox AI Window that is prompt-driven and relies on third-party cloud AI providers, producing machine-mediated summaries rather than direct browsing. While Mozilla frames AI features as something users can "turn off," the language suggests they may be enabled by default. The article questions whether this shift serves users or primarily supports Mozilla's business needs. Importance for marketers: If browsers mediate the web through AI summaries, direct brand experiences may be diluted. Marketers should expect more traffic to be "pre-consumed" by AI layers and double down on clarity, credibility, and brand cues.

You can find the previous issue of AI Update here.

Editor's note: GPT-5.2 was used to help compile this issue of AI Update.