Catch up on select AI news and developments from the past two weeks or so:
OpenAI declares 'code red' as rivals close in on ChatGPT. OpenAI CEO Sam Altman has reportedly declared a "code red," telling staff to urgently improve ChatGPT as competitors like Google and Anthropic catch up. The company is pausing new initiatives such as ads, shopping and health agents, and the Pulse personal assistant to focus on speed, reliability, personalization, and broader question coverage. Daily coordination calls and temporary team transfers will support the push. The move underscores intensifying competition and the challenge of funding massive AI investments while still figuring out profitability. Importance for marketers: ChatGPT's road map is shifting toward core quality rather than new surfaces. Expect rapid, sometimes disruptive, UX changes that affect integrations, content workflows, and any planning based on future ad or commerce features.
Google tests merging AI Overviews with AI Mode in Search. Google is testing a unified experience that connects AI Overviews with conversational AI Mode directly from the main search results page on mobile. Users who see an AI Overview at the top of results will be able to tap through and continue asking follow-up questions in an integrated chat with Gemini. Previously, people had to decide whether to start in classic search or switch into AI Mode. The experiment rolls out globally on mobile and aligns with Google's vision of frictionless, conversational information seeking. Importance for marketers: This further blurs the line between search and chat. Winning visibility means optimizing content so AI Overviews and conversational follow-ups surface your brand, not just ranking in traditional blue links.
ChatGPT adds shopping research to power product discovery. OpenAI launched a shopping research feature in ChatGPT that turns product discovery into a multistep conversation. Users describe what they need, answer follow-up questions about preferences and budget, and receive a customized buyer's guide. The system pulls price, availability, specs, reviews, and images from across the web, and users can refine results by marking items as "Not interested" or "More like this." The feature, powered by a shopping-specialized GPT-5 mini variant, currently performs best in consumer categories such as electronics, beauty, home, and sports. Importance for marketers: Early-stage product comparison can now happen entirely inside ChatGPT. Retailers and brands must ensure their product data, structured content, and site visibility align with how OpenAI's shopping system discovers, evaluates, and ranks items.
Retailers race to be seen by AI as holiday shopping goes agentic. Big US retailers are adapting to an AI-shaped shopping landscape, where tools like ChatGPT and Gemini recommend products alongside traditional search and ads. With most of the projected $253 billion in online holiday sales still coming through classic web journeys, AI referrals are small but growing and show higher purchase intent. Brands are experimenting with AI-only websites hidden from human shoppers, ramping up blog and Reddit content, and seeking coverage in editorial "best of" lists that scrapers ingest. Early data shows AI-driven traffic rising sharply year over year. Importance for marketers: Generative engine optimization is becoming operational reality. Teams need content, technical SEO, and reputation tactics designed explicitly to feed AI scrapers and agentic recommendation systems, not just human readers.
Ads appear headed to ChatGPT's free app experience. Code discovered in a beta version of the ChatGPT Android app suggests OpenAI is preparing an ads feature, with references to "bazaar content," "search ad," and a "search ads carousel." That implies the free ChatGPT app may soon include advertising alongside or within responses. OpenAI has so far relied on subscriptions and API usage for revenue, but IT remains unprofitable and reportedly faces large funding needs to sustain its growth. CEO Sam Altman has previously said he is not opposed to ads if they are implemented carefully. Importance for marketers: If search-style ads arrive in ChatGPT, it effectively becomes a new paid media channel. Marketers should expect auction dynamics, creative formats, and measurement questions similar to search—only now embedded inside conversational experiences.
DeepSeek releases open-source models that challenge GPT-5. Chinese startup DeepSeek unveiled two large models, DeepSeek-V3.2 and DeepSeek-V3.2-Speciale, claiming performance on par with or better than GPT-5 and Google's Gemini 3.0 Pro on elite math, coding, and reasoning benchmarks. A sparse attention architecture dramatically cuts inference costs and supports long context windows, enabling cheaper processing of large documents. Both models, with hundreds of billions of parameters, are released under an MIT open-source license, including full weights and tooling. Regulators in Europe and the US are already scrutinizing DeepSeek over data and national security concerns. Importance for marketers: Frontier-level capabilities are moving into free, open models. This could reduce dependence on a few US providers, but also raises questions about data governance, compliance, and where sensitive customer or campaign data is processed.
AI helps drive record $11.8 billion in Black Friday online sales. US Black Friday online spending hit a record $11.8 billion, up 9.1% from 2024, according to Adobe Analytics. AI-powered shopping tools played a major role, helping consumers navigate inflation worries, tariffs, and flat discount levels. AI-driven traffic to retail sites rose more than 800% compared with last year, aided by tools such as Walmart's Sparky and Amazon's Rufus. Globally, Salesforce estimates AI and agents influenced $14.2 billion in Black Friday online sales. Higher prices and stronger luxury demand offset fewer units per order. Importance for marketers: AI assistants now materially shape holiday commerce, from price discovery to basket composition. Brands should instrument journeys to understand AI-assisted paths to purchase and adjust promotion tactics, messaging, and merchandising accordingly.
EU plans to delay high-risk AI rules until 2027 after industry pushback. The European Commission proposed delaying enforcement of stricter AI Act rules on high-risk use cases from August 2026 to December 2027. The changes, part of a broader "Digital Omnibus," aim to streamline digital regulations, cut administrative burdens, and respond to criticism from Big Tech and other businesses. Affected areas include biometric identification, hiring and exams, utilities, road traffic, health services, creditworthiness, and law enforcement. Proposed GDPR tweaks would also allow large tech firms to use Europeans' personal data to train AI models under certain conditions. Importance for marketers: The delay gives organizations extra time to align their AI-driven hiring, customer scoring, and service workflows with EU rules. However, eventual compliance is inevitable, especially for profiling, risk scoring, and use of personal data for model training.
Study finds AI companies' safety practices lag global standards. A new edition of the Future of Life Institute's AI safety index concludes that major AI firms, including Anthropic, OpenAI, xAI, and Meta, fall far short of emerging global safety standards. An expert panel found none had robust plans to control advanced, potentially superintelligent systems, despite rising public concern about harms such as self-harm, psychosis, and AI-enabled hacking. The report criticizes both limited regulation and aggressive lobbying against binding rules, even as companies invest hundreds of billions in new AI infrastructure. Importance for marketers: As safety narratives intensify, brands deploying AI for customer engagement, personalization, or support must show credible guardrails. Clear policies, human escalation paths, and transparent messaging about AI limitations will increasingly factor into brand trust and corporate reputation.
White House pauses order that would preempt state AI laws. The Trump administration has paused a draft executive order that would have empowered the Justice Department to challenge state AI laws and threaten broadband funding to states that impose their own AI regulations. The draft envisioned an "AI Litigation Task Force" to argue that such state rules unconstitutionally burden interstate commerce. Lawmakers from both parties opposed the plan, calling it an attack on state efforts to address fraud, deepfakes, and child abuse imagery. Critics said it would prioritize AI industry concerns over consumer protections. Importance for marketers: The pause preserves a patchwork of state-level AI rules for now. Brands using AI in advertising, personalization, and content generation still need to track state regulations and avoid assuming a single federal standard will override local restrictions.
EU edges toward banning social media and AI chatbots for those under 16 who don't have consent. The European Parliament adopted a nonbinding resolution calling for a ban on social media for teens under 16 without parental consent, extending to AI chatbots due to mental health concerns. The move aligns Europe with Australia's stricter youth protections and follows French President Emmanuel Macron's criticism of addictive algorithms and "Wild West" platforms. Implementation details remain unclear, but the plan may rely on a digital identity wallet for age verification, requiring teens to re-verify regularly. Critics warn of a pervasive "digital checkpoint" infrastructure. Importance for marketers: Youth-focused campaigns in Europe could face dramatic access and targeting changes. Age-gating, consent workflows, and channel strategies for teen audiences will need to adapt to stricter verification regimes and possibly reduced reach.
EU opens antitrust probe into Meta's AI plans for WhatsApp. The European Commission has launched an antitrust investigation into Meta over policies that may restrict rival AI assistants on WhatsApp, giving an advantage to Meta's own Meta AI. Complaints from smaller companies alleged that changes to access rules could block competing chatbots from reaching users on the platform. EU officials are considering interim measures and note that the case is separate from, but complementary to, digital gatekeeper rules under the Digital Markets Act. Italy's antitrust authority is running a parallel investigation. Importance for marketers: If WhatsApp becomes more locked down around Meta's AI, access to users through third-party assistants could narrow. Brands betting on independent AI agents for messaging campaigns and customer service must track how platform owners control AI distribution and integrations.
Anthropic launches Claude Opus 4.5 for enterprise-grade coding and analysis. Anthropic announced Claude Opus 4.5, its newest flagship model, tailored for professional developers and knowledge workers such as analysts and consultants. Opus 4.5 follows recent releases of Sonnet 4.5 and Haiku 4.5 and is positioned as state-of-the-art for agentic coding tasks, outperforming rivals like Gemini 3 Pro and GPT-5.1 on SWE-bench Verified. Anthropic says the model excels at working with spreadsheets, slide decks, and deep research, and will serve as the default for its Pro, Max, and Enterprise offerings. Importance for marketers: Stronger coding and analysis capabilities make it easier to automate campaign reporting, dashboard building, and data exploration. Enterprise teams can offload more operational work to AI and redirect human effort toward brand, creative, and strategy.
Anthropic introduces Claude for Nonprofits with steep discounts and connectors. Anthropic, in partnership with GivingTuesday, launched Claude for Nonprofits to help resource-constrained organizations adopt AI. Eligible nonprofits can receive up to 75% off Team and Enterprise plans, with access to Sonnet 4.5, Haiku 4.5, and Opus 4.5 on request. The program includes open-source connectors to Benevity, Blackbaud, and Candid, plus a free AI Fluency for Nonprofits course via Anthropic Academy. Partners such as Bridgespan and Vera Solutions will support strategy and implementation. Importance for marketers: Nonprofit marketing and development teams gain access to enterprise-grade AI at lower cost. That can raise expectations for content quality, grant output, donor personalization, and reporting even in smaller organizations, reshaping benchmarks for nonprofit communications and fundraising.
Mistral debuts new large and small models to stay in the AI race. French AI startup Mistral released a new suite of models, including Mistral 3, which it calls the world's best open-weight multimodal, multilingual large model, and a compact Ministral 3 model for devices, robots, and drones. The company, recently valued at 11.7 billion euros after a funding round led by ASML and Nvidia, positions the models for agents, retrieval-augmented systems, and complex enterprise workflows. Ministral 3 can run on a single GPU, lowering costs and supporting on-device use without network access. Importance for marketers: Mistral's open-weight models expand options beyond US giants, especially in Europe. Flexible deployment on devices and lower-cost infrastructure could enable more localized, privacy-conscious AI experiences in campaigns and customer touchpoints.
Nvidia takes $2 billion stake in Synopsys to deepen AI design tools. Nvidia has invested $2 billion in chip design software maker Synopsys, acquiring about 4.8 million shares in a private placement and becoming one of its largest shareholders. The expanded tie-up will see Synopsys use Nvidia's developer tools and libraries to build AI-powered applications for chip design, verification, and other electronic design automation tasks. The deal follows a string of major AI investments by Nvidia, including stakes in OpenAI and Intel. Synopsys shares rose on the news while Nvidia's dipped slightly. Importance for marketers: Nvidia's growing influence across the AI hardware and software stack could reinforce its position as the default infrastructure layer. Over time, this may affect model availability, pricing, and performance for SaaS tools and marketing platforms your teams depend on.
ByteDance launches Doubao-powered AI voice assistant on ZTE phones. ByteDance is introducing an AI voice control tool, powered by its Doubao large language model, that will debut on ZTE's Nubia M153 smartphone before rolling out to other manufacturers. Users will be able to find content and book tickets via voice, competing with AI features from Huawei and Xiaomi. Apple's Apple Intelligence is not yet available in China, though Alibaba has announced collaboration plans. ByteDance stresses it will not build its own phones and is partnering across the ecosystem. Importance for marketers: In China, voice AI entry points are diversifying beyond a single platform. For brands targeting Chinese consumers, discoverability strategies must consider device-level assistants and super-app ecosystems, not just search or social.
Wikipedia seeks more AI licensing deals to cover scraping costs. Wikipedia co-founder Jimmy Wales says the nonprofit is negotiating additional licensing agreements with AI companies, similar to its existing deal with Google, to offset infrastructure costs from heavy AI crawling. High-volume automated access forces Wikipedia to pay for more servers and memory, even though its content remains free to individuals. Wales argues donors did not intend to subsidize multibillion-dollar AI products, and suggests technical tools like Cloudflare's AI Crawl Control may limit bots if companies do not pay. Importance for marketers: As key knowledge sources demand compensation or restrict access, training data and content availability may shift. Brands that rely on reference-style content, UGC, or wikis should anticipate changing terms for how AI systems can use and quote that material.
Amazon workers warn about an all-costs-justified AI strategy. More than 1,000 Amazon employees have anonymously signed an open letter criticizing the company's "all-costs-justified" AI push. Organized by Amazon Employees for Climate Justice, the petition warns that AI investments risk fueling layoffs, environmental damage from energy-hungry data centers, and abuse in surveillance and deportation contexts. Workers say executives overstate AI benefits while downplaying internal dissatisfaction and climate impact. Signers include senior engineers, product leaders, marketers, and warehouse staff, many of whom fear retaliation in a more restrictive internal climate. Importance for marketers: Internal resistance to AI rollouts can affect brand reputation, recruiting, and customer trust. Marketers must factor employee sentiment and social responsibility into how they communicate AI initiatives, both externally and in employer branding.
Amazon pulls AI anime dubs after fan and creator backlash. Amazon's Prime Video quietly rolled back a beta program that used AI to generate English and Spanish dubs for several anime titles; the move came after fierce criticism from fans and voice actors. The AI dubs, seen as undermining human performers and delivering poor quality, triggered online outrage, including high-profile critiques from well-known dub actors. Some English AI dubs have been removed, though certain Spanish versions remain. The episode follows broader industry hesitation around visible generative AI use in films and television. Importance for marketers: Audiences can strongly reject AI-generated creative work if it feels like cost-cutting at the expense of quality or livelihoods. For brand content, transparency, creative standards, and collaboration with human talent are essential to avoid reputational blowback.
OpenAI to acquire Neptune for deeper training analytics. OpenAI has agreed to acquire Neptune, a startup that builds monitoring and debugging tools for AI model training workflows. The two companies already collaborated on a metrics dashboard for foundation-model teams. Neptune will wind down its external services, and terms of the deal were not disclosed. OpenAI plans to integrate Neptune's fast, precise analysis system into its training stack to expand visibility into how models learn. The acquisition joins a string of OpenAI deals aimed at strengthening internal tooling and capabilities. Importance for marketers: Better training analytics can translate into more reliable, capable models powering the tools marketers use daily. Over time, that may mean fewer unexplained failures, more consistent output quality, and faster iteration on new features in marketing and creative platforms.
Russian tech leader compares AI power to a new nuclear club. Alexander Vedyakhin, first deputy CEO of Russia's Sberbank, argues that AI will create a "nuclear club" of nations with their own large language models, conferring influence comparable to nuclear technology. He says Russia must maintain at least two or three homegrown models for sensitive domains and cannot rely on foreign systems. Vedyakhin claims Sberbank's GigaChat models rival top Western systems but warns that energy and infrastructure demands make AI investment risky and potentially unprofitable, especially amid sanctions and limited hardware access. Importance for marketers: Geopolitical competition around AI may drive national preferences and restrictions on which models can be used in certain markets. Global brands should anticipate localization requirements, data residency rules, and political sensitivities around AI vendor choice.
Trump's Genesis Mission aims to use AI for scientific and energy breakthroughs. President Trump signed an executive order launching the "Genesis Mission," billed as the largest mobilization of US scientific resources since Apollo. The Department of Energy will build an AI-enabled platform to support research in areas such as energy, national security, and drug discovery. Officials say AI will automate experiments, speed simulations, and generate predictive models to cut energy costs and accelerate breakthroughs. The initiative has no public cost estimate but promises new partnerships with companies, including Nvidia and Dell. Importance for marketers: Large government-backed AI programs can accelerate innovation in energy, healthcare, and infrastructure, indirectly creating new markets and narratives. Marketers in B2B and public sector contexts should watch where funding flows, as it will shape future demand, partnership opportunities, and policy direction.
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.
