Welcome to Week In Review
From new tools transforming creative work, to campaigns reshaped by automation, we track the biggest shifts at the intersection of AI, media, and marketing. This week, OpenAI brings product carousels to ChatGPT ads, Article 50 of the EU AI Act comes into force, and Pubmatic launches guardrail architecture inside AgenticOS.
Top Stories of the Week
OpenAI Brings Product Carousels to ChatGPT Ads
OpenAI has launched product carousel ad formats inside ChatGPT, allowing retailers to display multiple products with images, prices and purchase links within a single AI-generated response. This mirrors a format long used by Google Shopping and Meta’s catalogue ads, and marks a significant step in making ChatGPT a genuinely shoppable surface.
Why it matters:
Product carousels are not a minor format update – they are the moment ChatGPT stops being an experimental ad surface and starts competing directly with Google Shopping and Meta for retail budgets. Until now, ChatGPT advertising has been primarily text-based, which limited its appeal to brand and consideration campaigns. Visual product carousels with prices and purchase links are the format performance marketers actually spend against, because they connect directly to conversion.
EU AI Act Article 50 Transparency Obligations Enter Into Force
From August 2, 2026, providers and deployers of certain AI systems must comply with the transparency obligations set out in Article 50 of the EU AI Act, with non-compliance triggering fines of up to €15 million or 3 percent of worldwide annual turnover. Under the new rules, chatbots and other interactive AI systems must tell users they are dealing with AI, not a human; deepfakes must be labelled; and AI-generated or altered content must carry machine-readable marks so it can be detected more easily.
Why it matters:
Article 50 is the first article in the Act to be enforced, and it lands immediately on every business running a chatbot, AI customer service agent, or conversational tool in the EU. Marketers will now need to ensure every AI-powered interaction their brand runs in the EU tells users they are talking to AI. The broader implication is that the EU has now begun active enforcement rather than simply legislating, and the December deadline for GenAI content marking is the next article marketers will need to plan for.
PubMatic Launches Guardrail Architecture for Autonomous Ad Buying
PubMatic has launched a five-step governance layer inside AgenticOS, its autonomous buying platform, restricting what AI agents can transact and recording every decision they make. The framework has five components: platform-wide and buyer-configured boundaries, inherited business rules, pre-approved asset libraries, authenticated approval workflows using one-time passcodes, and continuous audit trails with drift detection.
Why it matters:
The competitive ground in agentic advertising is shifting toward control. Agencies buying autonomously carry exposure they did not carry when a human trader signed off on each placement. PubMatic’s guardrail architecture is a direct response to the trust gap that has slowed agentic advertising adoption.
Quote of the Week
Brands & Agencies
Jellyfish Launches AI Ads Optimisation for “Share of Model”
BrandTech’s Jellyfish has launched an AI-powered ads optimisation product designed to help brands improve their “share of model” – the frequency and quality with which they appear in AI-generated answers and recommendations – connecting paid media optimisation directly to AI visibility outcomes for the first time.
Fossil Uses AI to Identify Target Profiles, Earns 58.8 Million Impressions
Marketing Dive reported this week that Fossil Group – which owns brands such as Fossil and Skagen – used AI to identify and target high-value consumer profiles for a recent campaign, generating 58.8 million impressions through more precise audience modelling than traditional demographic targeting would have achieved.
Only 53% of Marketers Are Seeing Meaningful ROI From AI, TransUnion Finds
A new TransUnion study titled AI Confidence-Readiness Paradox found that just over half of marketers report meaningful return on their AI investments, with the gap between adoption and impact widening as the pressure to deploy AI increases faster than the measurement frameworks to evaluate it.
IAB Releases “Measuring Visibility in the AI Era” Framework
The IAB has released a new industry framework called Measuring Visibility in the AI Era to help brands, publishers and agencies navigate AI-powered discovery, setting out how to define, measure and optimise for visibility in AI-generated answers and recommendations as the metric increasingly replaces traditional search ranking as the primary measure of digital discoverability.
Media
News Corp CEO Calls Out “Slimy AI Slop” as Australian Subscriptions Fall
News Corp reported its EBITDA news division experienced a fall in subscriptions, with CEO Robert Thomson using the results to call out “slimy AI slop” publishers and unnamed AI companies for training on and reproducing editorial content without compensation – framing the subscription decline in the context of AI-driven traffic disruption rather than editorial performance.
Director James Gray Says AI Is Killing Hollywood’s Soul
Film director James Gray – who directed films Ad Astra and the Lost City of Z – has publicly argued that the industry’s rush to adopt AI is stripping Hollywood of the creative risk-taking and human artistry that define great filmmaking – positioning himself alongside del Toro and other directors who have drawn a firm public line against AI in production.
Eli Roth Uses AI-Generated Shots in Ice Cream Man
Director Eli Roth – who directed films Hostel and Cabin Fever – disclosed this week that several shots in his new film Ice Cream Man were created using GenAI, becoming one of the more prominent directors to acknowledge AI production use in a mainstream release.
Nvidia Creates AI Video Detector
Nvidia has launched an AI-powered tool capable of detecting whether video content was generated by AI, entering the content provenance and authenticity detection market at a moment when the volume of AI-generated video is accelerating and regulatory pressure to label this content grows.
Tech
Snapchat Offers Third-Party AI Ad Integration
Snapchat has opened its advertising platform to third-party AI integrations through an Model Context Protocol (MCP) server, allowing external AI tools to connect to its ad systems and automate elements of campaign creation, targeting and optimisation — extending the agentic ad-buying trend into one of the most youth-skewing social platforms.
TikTok Rolls Out Longer-Form AI Video Ads
TikTok has expanded its AI video ad capabilities to include longer-form formats, enabling advertisers to generate AI-produced video ads beyond the short-form lengths previously available.
Google’s Deceptive Practices Fuel Tensions With SSPs and Publishers
Adweek reported this week that Teads has made allegations of deceptive practices against Google, adding to growing tension between the search giant and independent SSPs and publishers over how Google’s ad systems allocate inventory and pricing across the open web.
Cloudflare Launches Wallets and CloudflarePay to Enable Secure AI Agent Commerce
Cloudflare has launched a payments infrastructure layer – including digital wallets and CloudflarePay – designed to allow AI agents to complete commercial transactions securely, without human intervention at the point of payment. The launch positions Cloudflare as infrastructure for the agentic commerce era, moving beyond its web security origins into the financial rails of the AI economy.
Startup of the Week
Gravity, an adtech startup that places ads inside AI chatbots, raised a $30.5 million Series A co-led by Lightspeed Venture Partners and Committed Capital, bringing total funding to $38.5 million. The company runs a full-stack ad platform: a demand-side platform for buying ads within AI chatbots and agents, a supply-side platform for developers to monetise their AI apps through ads, and an exchange connecting the two.
Number of the Week
Gartner claims that AI will operate “behind the scenes” to determine which ads are shown, to whom, and at what cost, distinct from the generative AI used to create ad creative.
