LLMs

Reddit is surging with advertiser demand; not everyone is Substack material; and Trump released his AI Action Plan.

Carat is using “people-powered” AI to build agents that understand how to best reach specific audience personas.

Amazon acquires AI-equipped wearable manufacturer Bee; the UK’s CMA shares competition guidelines for Google and Apple; and AI models may be learning from each other in unexpected, potentially harmful, ways.

Pixels attached to articles explaining a recent health diagnosis – without consent  – led Healthline to a record $1.55 million fine for violating CCPA. Plus: the new AI contract.

The IAB Tech Lab’s new initiative suggests regulations for how AI bots can access content, ensuring that publishers are fairly compensated.

Sigma, MiQ’s new AI-powered ad platform, gives advertisers better analytics and attempts to unify the fragmented data landscape.

TikTok isn’t quite such a success for luxury beauty brands anymore; expect even more tariff uncertainty; and LLMs have a hard time with “the taxonomy problem.”

Jerry Dischler leaves Google; a bunch of marketing execs join AI companies; agency holdcos don’t know which way is up.

AI is being embedded in tools, tagged in decks and tossed into internal sprints. But for agencies, publishers, brands and platforms, there’s still no shared definition of success.

Threads will introduce ads to capitalize on users fleeing X; Perplexity tests ads and sponsored queries; and Amazon pulls the plug on Freevee.