Daily AI Briefing

Google's school AI opt-out comes with a catch

Google will switch off its AI tools for K–12 accounts by default early next year, leaving it to administrators to turn them back on. But here’s the catch: schools that block AI entirely also lose student access to Google Search, because its AI features can’t be disabled separately.

The move follows worries about cheating and students’ well-being, plus pressure from lawmakers. Google’s also pushing Guided Learning, which coaches students through problems step by step. The company is piloting the tool in Search at select U.S. schools; it’s already in Classroom globally.

An opt-out that costs you Search isn’t much of an opt-out.

More coverage:

• TechCrunch: Google will let schools disable its AI tools, but not without sacrificing Search

An MBA taught by AI avatars

San Francisco Bay University is using AI avatars to deliver lectures in its new online MBA Pro+ program, with a chatbot tutor named Bayley on hand for guidance. The degree costs $10,000, targets working and nontraditional students, and has 27 enrollees so far. Human faculty still hold virtual office hours and help shape the course materials.

It’s part of a bigger bet. SFBU rebuilt itself around AI after enrollment dropped to 30 students in 2023 — it’s now around 400. The school says AI extends faculty support and makes courses more flexible. The avatars, though, raise a fair question: what does an instructor’s job look like next?

The $1.5 million AI engineer

Companies are hiring forward-deployed engineers (FDEs) to build AI in-house and get projects out of pilot purgatory, putting them in direct competition with tech firms and consultancies. The elite ones earn $586,000 to $1.45 million a year, according to a six-month study by executive search firm Christian & Timbers. Of the roughly 17,000 FDEs in the U.S., only about 2,000 make the elite cut — and around 80% of those are current or former Palantir engineers.

Target, Barclays, Bausch + Lomb and American Express have all posted FDE roles. With executives under pressure to show real returns on AI, one recruiter expects demand to stay hot through at least 2028.

Turns out the hardest part of AI isn’t the model. It’s the plumbing.