Two AI answers with real consequences were reported yesterday, an illegal gambling recommendation and a wrong hunting-season date that ended in a citation, while Google, Shopify and LiveRamp each extended what AI agents can do next.

Confident and Wrong: When the Answer Carries a Consequence

Two AI answers reported yesterday did not just underperform. They produced a real citation and a real safety risk.

AI Mode Keeps Expanding: From Watching the Web to Booking the Room

Google widened AI Mode’s reach on two fronts, a monitoring feature and a hotel booking standard, without locking down the details of either.

Agents Move From Browsing to Buying

Two platforms extended AI agents further into the purchase itself, one at checkout and one inside the ad stack.

Rules and Rankings: An Appeal in Brussels, a Spike Nobody Called Final

One story of oversight is legal, the other algorithmic, and both are still unresolved.

Today’s Quick Hits