Four platforms, one very AI kind of week.
Say cheese: Meta just launched Muse Image, its first in-house image model, live in Meta AI now.
It’s a prompt-driven creative engine that renders legible text and even functional QR codes. The best part? It’s coming to advertisers through Advantage+ creative.
Know your audience: OpenAI is rolling out audience list uploads on its new ChatGPT Ads platform.
You can feed it raw or hashed emails and phone numbers, giving you real first-party targeting on an emerging channel. Early days, but worth a look.
Now for the cleanup crew: Reddit has a bot problem. And it says its upgraded AI defenses are cutting spam exposure and revoking nearly 2M fake votes daily.
Suspicious automated accounts now get asked to verify their humanity. If you lean on automation for Reddit, expect a tighter, more human-vetted room.
And a legal plot twist: A German court ruled Google can be held liable for false claims its AI Overviews generate, treating them as Google’s own statements, not third-party links.
If that view holds, Google may be pushed to source and link out more to dodge liability, which could mean better attribution for publishers.
It’s anything but smooth sailing with AI developments…
