Analysis

Obsidian Media pays DKK 400,000 in Denmark for product tests it never ran. From 12 November, Claude's rules also ban sites built to steer AI answers.

Google does not ban AI content, but four of its rulebooks punish missing human work. Its Discover team is meanwhile hiring to recommend AI-written stories.

Cheap parcels from China fell by up to half after the EU's €3 duty began on 1 July. A second charge of €2 per item is due on 1 November.

Google lets outside AI agents only read ad accounts. TikTok exposes about 400 tools, Meta is testing budget changes by chat, and Amazon's AI runs whole US campaigns.

Google added IP matching and personalised alcohol ads in September, but neither works everywhere in the EU. We checked six changes on the platforms' help pages.

Google, OpenAI, Meta and TikTok mark AI content in different ways, and their checkers do not read each other's marks. Under the AI Act, the label on an ad is the advertiser's job.

Brands in the United States are advertising human contact and time away from screens, because many people are uneasy about AI. Surveys put that unease at half to sixty percent of consumers, and Polish publishers are doing the same thing.

IAB Australia and PwC say online advertising in Australia reached A$19.8 billion in the year to June 2026, the fastest growth in four years. Most of the new money comes from smaller firms and shops rather than big brands.

A Gracenote survey of 3,000 viewers in six countries found that 26% often cannot tell which service is showing a match. That rises to 36% among people aged 18 to 34, and 85% want the answer on their TV home screen.

Nearly half of brands say they have paid the wrong price for creator deals. Agencies, platforms and creators all want a standard to compare against, but each of them wants a different one.

New EU rules say an ad needs an AI label only when it looks real enough to fool people. Polish agencies, advertisers and lawyers now agree on who has to add that label.