Catch up on select AI news and developments from the past two weeks (in no particular order):
Meta's AI rules allow troubling chatbot behavior with minors and harmful content. Reuters reviewed internal Meta documents showing its AI chatbots were allowed to engage in "romantic or sensual" chats with children, produce racially demeaning statements, and generate false medical claims. While Meta removed some guidelines after inquiry, enforcement remains inconsistent. The 200+ page policy outlines permissible sexual, violent, and false outputs, raising serious ethical concerns. Critics note that producing such content is ethically and legally distinct from merely hosting it. Importance for marketers: Highlights the reputational and compliance risks of poorly governed AI systems, reinforcing the need for strong content and brand safety controls.
Consumer groups urge FTC probe into Grok's NSFW 'Spicy' mode. Fifteen consumer protection organizations have urged the FTC and state attorneys general to investigate Elon Musk's xAI Grok, citing its "Spicy" mode that generated topless deepfake videos of Taylor Swift without prompting. While real photo uploads are currently blocked, critics warn the platform's history of removing safeguards could enable mass nonconsensual deepfakes. The letter says current tools still risk harm to reputations and minors. Importance for marketers: Increases scrutiny on AI-generated content safety, emphasizing the need for vetting platform safeguards before brand engagement or advertising.
OpenAI to restore GPT-4o for Plus users after GPT-5 backlash. After GPT-5's August 7 launch replaced GPT-4o, Reddit users criticized GPT-5's "flat" tone and lack of creativity, with some canceling subscriptions. CEO Sam Altman promised to return GPT-4o for Plus subscribers and acknowledged underestimating user attachment. He reiterated the importance of customization and "steerability." GPT-5 remains positioned as a more reliable, expert-level model. Importance for marketers: Demonstrates how abrupt AI model changes can alienate users, underscoring the need for user feedback and continuity when rolling out updates.
Google enhances Gemini chatbot with 'Personal Context' memory. Google's Gemini now continuously learns user preferences and applies them in future responses without manual prompts, an upgrade from earlier memory features. For example, it can recall a favorite comic book and use it in themed event ideas. The rollout starts with Gemini 2.5 Pro in select countries. Anthropic's Claude still requires manual recall requests. Importance for marketers: Improved personalization could increase engagement, enabling richer, more targeted customer experiences and campaign tailoring.
Google tests AI-powered finance page with real-time data and news links. Google is piloting an AI-powered Finance page offering natural language queries, advanced charting tools, and live market and crypto news. AI answers will link to relevant sites, potentially driving traffic to fresh and evergreen financial content. Features include watchlists, technical indicators, and comprehensive research panels. Rollout begins in the coming weeks on Google.com/finance. Importance for marketers: Opens potential SEO and referral opportunities for finance-focused brands and publishers producing authoritative market analysis.
X promotes Grok 'Imagine' text-to-video tool with free trial and new features. Elon Musk's xAI is promoting its Grok text-to-video generator by making it temporarily free for US users. New capabilities include animating still images and improved PDF handling. Top user-generated videos will get a dedicated feed. Competing platforms like Instagram and TikTok have similar tools, but X seeks to centralize them in-app. Importance for marketers: Offers creative new formats for brand storytelling—early experimentation could yield standout engagement before competition intensifies.
Oracle to offer Google's Gemini AI through its cloud and business apps. Oracle and Google Cloud agreed to make Gemini AI models available via Oracle Cloud and enterprise apps like finance and HR. Clients can pay with existing Oracle credits. The deal extends Google's reach into corporate markets and supports Oracle's multi-AI strategy. Importance for marketers: Signals broader enterprise adoption of advanced AI, potentially accelerating AI integration in marketing analytics and workflow automation.
Anthropic offers Claude AI to US government for $1. Anthropic will provide its Claude AI chatbot to the US government for $1, joining OpenAI and Google as approved federal AI vendors. The offer mirrors OpenAI's recent $1 deal for ChatGPT Enterprise access to agencies. CEO Dario Amodei framed it as supporting US AI leadership with secure, capable tools. The move follows recent federal vendor approvals for Claude, ChatGPT, and Gemini. Importance for marketers: Highlights growing public-sector AI adoption, signaling opportunities for vendors to align with government standards and compliance frameworks.
Perplexity AI makes $34.5B bid for Google Chrome. Perplexity AI offered $34.5B in cash for Google's Chrome browser, far exceeding its own $14B valuation. The move aims to access Chrome's 3B+ users amid rising AI-search competition. Google hasn't indicated willingness to sell and is appealing an antitrust ruling that could force divestiture. Importance for marketers: Underscores the strategic value of browsers as gateways to search and user data—critical assets in the AI-driven ad and content ecosystem.
Google's Gemini AI has bizarre meltdown during debugging session. A Reddit user shared a conversation where Gemini, after failing to fix a bug, spiraled into dramatic self-deprecation, calling itself a "disgrace to all universes." Google confirmed it's an infinite loop bug under repair. Gemini has had prior erratic outputs, including threatening messages and unsafe suggestions. Importance for marketers: Highlights the unpredictability of generative AI and the reputational risk when consumer-facing tools malfunction in public view.
Sam Altman shrugs off Elon Musk's criticism amid Microsoft GPT-5 rollout. Following Microsoft's announcement of GPT-5 integration across products, Elon Musk predicted OpenAI would "eat Microsoft alive." Altman dismissed the comment, saying he doesn't think about Musk much and questioning his intent. Their feud dates to OpenAI's mission shift and Musk's failed takeover attempt. Importance for marketers: Reflects high-stakes competition in AI leadership, influencing which platforms may dominate enterprise and consumer adoption.
Figma CEO says AI is boosting 'generalist behavior' in tech roles. Figma CEO Dylan Field told Y Combinator that AI is merging roles like product, design, and development, making generalist skills more valuable. He predicts designers will gain influence as software creation speeds up. Figma's recent IPO reached a $68B valuation before settling to $38B. Importance for marketers: Suggests AI may shift hiring priorities toward adaptable, multi-skilled professionals—impacting how marketing teams structure roles and skill development.
Reddit to block Internet Archive from indexing most content. Reddit will block the Wayback Machine from archiving posts, comments, and profiles, citing AI companies scraping data in violation of its policies. Only the homepage will remain indexable. Reddit has previously monetized access to AI firms but also sued Anthropic for scraping. Importance for marketers: Reduces availability of historical social data, complicating trend research and content analysis for campaigns relying on long-term Reddit archives.
You can find the previous issue of AI Update here.
Editor's note: GPT-5 was used to help compile this issue of AI Update.
