Digital news to watch: Google admits AI Search reports can be misleading

5 min read
AI Search reports

In this week’s digital news to watch, Google’s EU search layout overhaul reshapes visibility for European marketers, while a separate Google admission reveals its AI Search reports have been unreliable all along, quietly undermining performance data everywhere.

Amazon is piloting ads inside ChatGPT via its DSP, opening a new front in AI surface advertising. Elsewhere, Google Ads promotional credits are being clawed back after advertisers already spent to earn them, TikTok blocked a Meta ad campaign aimed at pressuring rivals into its child safety settlement, and Meta launched its Muse AI agent in a bid to close the gap on OpenAI, Anthropic and Google.

Google changes European search layout due to EU laws

Google has altered how search results appear across Europe to comply with strict European Union tech regulations. When people search for commercial services like flights, hotels, or products, Google now displays dedicated sections specifically for price-comparison websites and direct suppliers. While the EU hopes this creates a fairer playing field for smaller competitors, Google argues it will make searching less convenient and could reduce direct bookings for individual businesses.

For website owners and marketers, this means that standard search rankings are being pushed aside in favour of these government-mandated layout boxes. Companies operating in Europe have to adapt quickly to these changes, as their visibility now depends heavily on whether they are listed in these specific comparison blocks rather than traditional search results.

Google admits its AI Search reports can be misleading

Google has acknowledged that Google Search Console does a poor job of measuring performance in AI-powered search results. Because the system relies on older counting rules, it can accidentally exaggerate how many people actually saw a website inside an AI summary, while missing visits that are hidden behind “read more” buttons. Furthermore, every link inside an AI answer gets lumped with the score of the whole box rather than showing its true individual position.
This means business owners and marketers cannot fully trust their current AI dashboard numbers when trying to measure success. Relying blindly on these flawed reports can lead companies to think they are performing better in AI search than they really are, forcing them to look at alternative metrics to understand their true online visibility.

Amazon pilots ad services in ChatGPT

Amazon advertisers can now extend campaigns into ChatGPT conversations, bought through Amazon DSP as a managed service, with OpenAI controlling ad delivery on its own systems. Delta Vacations is among the select US brands piloting the integration. Ads run as sponsored text and image units below organic responses, priced on CPC or CPM, US-only for now. It comes weeks after ChatGPT ads hit a $1 billion annualized revenue run rate, and follows smaller retail media tie-ups with Target and Albertsons, though Amazon’s is by far the biggest partner yet. OpenAI had reportedly also discussed a deal with The Trade Desk before choosing Amazon.

Google Ads credits pulled after advertisers already spent to qualify

Search Engine Journal reports two cases where Google Ads promotional credits were invalidated after advertisers had already spent the required amount to earn them, in one case a $3,200 credit marked “Invalidated” over a month later. He says there’s no clear appeal path, and once the spend has happened it can’t be clawed back. Melamed also argues credits push advertisers to bid more aggressively, inflating auction prices for everyone, though he offers no evidence for that. Google’s Ads Liaison Ginny Marvin acknowledged the complaint on LinkedIn but gave no explanation or fix. Takeaway: check promotional credit eligibility terms closely rather than treating the credit as guaranteed budget.

TikTok blocks Meta’s pressure campaign

TikTok rejected Meta ads urging TikTok and YouTube to join its child safety settlement, ruling the ads counted as prohibited political content. Meta agreed to pay up to $16.7 billion to settle US state claims over Facebook and Instagram harming children, but roughly $5 billion of that hinges on TikTok and YouTube signing up to similar restrictions and payments. Meta argues that if only it faces new limits, teens will simply migrate to rival apps. TikTok and YouTube haven’t publicly responded, and both reportedly withdrew from an HHS panel on children’s screen time after appearing on a draft agenda. The rejection shows Meta can’t force the issue through paid media, so the settlement’s wider impact now depends on backchannel pressure rather than ads.

Author

Stay in the loop
Share post

hi

Other posts you might like

A complete guide to Local SEO

A complete guide to Local SEO

TLDR: Local SEO optimises your online presence, so you show up when people search for a service near them. It
ModoFest 2026 💃

ModoFest 2026 💃

Modo25 has done it again, our 7th annual ModoFest was the event of the year, packed with brilliant talks, great

Popular topics

[other_categories]