Meta’s bold push to fully automate ad creation with AI
State of Play: Meta is working to let advertisers create entire ad campaigns, from concept to targeting, using AI alone by the end of 2026. Building on existing automation tools, the company aims to generate complete ads from just a product image and budget, then personalize them in real time based on viewer context like location.
Why it Matters: With over 97% of Meta’s revenue tied to advertising, full AI automation could redefine how brands reach customers. But the shift also raises concerns: brands worry about losing creative control, and quality issues still plague many AI-generated visuals. Meta’s move signals a coming transformation in digital marketing where humans may steer, but AI drives. AI in advertising is a hot topic, as the creative field was often perceived to be safe from automation; AI is unlikely to replace Cannes Lions-winning works, but do expect to see more of it in the generic call-to-action driven ads that flood the internet.
Google’s Veo 3 pushes AI video to the edge of reality – and trust
State of Play: Google has launched Veo 3, its most advanced AI video generator to date, a tool capable of producing hyper realistic video clips with cinematic visuals, synchronized dialogue, and sound effects, while previous AI tools were limited by visual glitches or surreal artifacts. Announced at Google I/O and available to premium subscribers for $249/month, Veo 3 is being hailed as a breakthrough in AI filmmaking. From eerily lifelike monologues to full-blown short films, users are flooding social media with Veo-generated clips that are often indistinguishable from real footage.
Why it Matters: Veo 3 marks a turning point in synthetic media. While Veo 3 is being embraced by some filmmakers for its creative potential, Veo’s near-flawless fidelity brings deepfakes to a new level of plausibility. In tests by journalists and watchdogs, Veo 3 has also shown it can generate realistic, misleading content. While Google has added watermarking and safety filters, experts warn that the tool’s outputs are already circulating online in ways that could spark real-world harm.
LinkedIn debuts first impression ads
State of Play: LinkedIn has announced the release of a new advertising format called First Impression Ads – a full-screen, vertical video format designed for single-day campaigns that ensures that the advertiser’s content is the first ad that a target member sees on a given day. This rollout comes alongside other advertising features (see our Platform Updates below for more detail).
Why it Matters: Long the natural home for corporate communications, this format may well make advertising on the platform even more appealing. From a corporate perspective, when it’s essential to create a moment around a major time-sensitive announcement, as opposed to an ongoing campaign, this format would increase its impact, clarity and recall among key audience groups, through its combination of full-screen coverage as well as first impression guarantee.
Trump to delay TikTok ban for a third time
State of Play: President Donald Trump has issued an executive order last week extending the deadline for TikTok’s forced divestment by another 90 days, allowing the app to continue operating in the U.S. This marks the third extension of a mandate passed by Congress last year requiring Chinese parent company ByteDance to sell its stake or face a nationwide ban. While the administration had previously been close to brokering a deal, Beijing pulled its approval following renewed U.S. tariffs in April.
Why it Matters: TikTok’s future in the U.S. remains entangled in both trade diplomacy and campaign strategy. Although Trump once called TikTok a national security threat, he has since softened his stance – especially after leveraging the platform to engage young voters in the 2024 election. The extension highlights how the app, once on the chopping block, has evolved into a political and cultural flashpoint. As trade tensions with China simmer and TikTok’s influence continues to grow, the delay underscores the balancing act between digital sovereignty, diplomatic friction, and domestic political calculus.
Reddit sues Anthropic for using user data without licensing agreement
State of Play: Reddit has filed a lawsuit against AI company Anthropic, claiming the startup repeatedly scraped its user-generated content, over 100,000 times, without permission to train AI models like Claude. The suit follows Anthropic’s refusal to enter a licensing deal, in contrast to Reddit’s paid agreements with companies like Google and OpenAI.
Why it Matters: As AI companies rush to secure training data, this legal clash highlights growing tensions over the value and ownership of online content. Reddit’s suit reflects a broader shift as platforms seek to monetize and protect their data, while calling for stronger safeguards around user privacy and consent in the AI arms race.
