30+ sources. Zero spin.
Cross-referenced, unbiased news. Both sides of every story.
Meta's AI Spending Spree Hits Three New Walls: Employee Surveillance Blowback, $9M School Settlement, and a Hardware Bet Nobody Asked For

The Surveillance Tool Nobody Can Opt Out Of Is Now a European Problem
Back in April 2026, Meta quietly launched something called the Model Capability Initiative (MCI) — a program that captures mouse movements, keystrokes, clicks, and occasional screenshots from US employees' company devices. No opt-out. The stated goal: train AI models to navigate software the way humans do.
Reuters obtained internal Q&A documents Meta provided to employees. Those documents confirm MCI pulls data from more than 200 apps and websites. More importantly, Meta acknowledged in writing that the tool captures the contents of emails and direct messages sent to US employees — regardless of where the sender or recipient is located.
"If a US-based colleague has the tool enabled while gchatting or emailing with someone outside the US, that activity would be captured," Meta wrote in the document, according to Reuters.
Meta spokesperson Dave Arnold told Reuters the company "carefully considered and mitigated potential privacy risks" and is "committed to complying with applicable laws and regulations."
Kayne McGladrey, a senior member of IEEE, told TechTarget: "This is something that can be done because we don't have a federal privacy act in the United States, whereas in other countries, this would be completely unacceptable."
Legal experts told Reuters that even limited capture of EU employee data "could put Meta in violation of the EU's General Data Protection Regulation (GDPR) rules." Under GDPR, companies need a valid legal basis for collecting personal data and must disclose what's being collected. Meta's framing — that it only notified non-US employees rather than obtaining consent — is almost certainly insufficient under that standard.
Meta has already been fined billions by EU regulators over data practices. A new GDPR enforcement action would come on top of that.
Employees have been complaining that MCI is consuming so much data it's burning through their home internet monthly quotas in days. Some employees distributed physical flyers asking colleagues to sign a petition against the program. The concern includes privacy and the fact that they're training their own replacements, following a layoff that cut roughly 10% of Meta's global workforce.
CEO Mark Zuckerberg has said publicly that AI systems will eventually handle much of the work in the tech industry. Employees are making that connection themselves.
Meta Just Paid $9 Million to a Kentucky School District
Bloomberg reported the financial details of a settlement between Meta, Snap, TikTok, YouTube, and the Breathitt County school district in Kentucky.
The breakdown, from documents obtained under Kentucky's open records laws:
- Meta: $9 million
- Snap: $8 million
- TikTok: $8 million
- YouTube: just over $2 million
Breathitt County had originally sued for $60 million to fund mental health programs. The district's principal told the court that most of his time is now spent dealing with "social media stuff" — students filming fights and bullying each other online.
This case was supposed to be a bellwether — an early test case to set the trajectory for over 1,300 similar lawsuits filed by school districts across the US. The companies settling before trial removes that data point, but the underlying problem remains. Bloomberg reports the full wave of lawsuits could ultimately cost these companies $400 billion. Breathitt County is a small rural district. Bigger districts mean bigger payouts. The first actual trial involving a school district is expected in February 2027.
Meta paid the most of any defendant — the platform most blamed for the addiction and mental health harms at the center of these cases.
Now Zuckerberg Wants to Sell You an AI Pendant
Meta is developing an AI pendant — a wearable that listens and records your conversations throughout the day. This comes from Meta's 2025 acquisition of Limitless, the startup that made a device called "Pendant" that did exactly that. Limitless CEO Dan Siroker described it as part of Meta's "vision to bring personal superintelligence to everyone."
Meta VP for wearables Alex Himel wrote in an internal memo that the company plans to release up to four new smart glasses models before the end of 2026:
- Modelo — launching as soon as June
- Luna — this fall
- RBM2 Refresh — also this fall (sounds like another Ray-Ban collaboration)
- Mojito VIP — December
Additional models named "Artemis" and "SSG" (supersensing glasses) are in testing for future release.
The goal: sell 10 million wearables in the second half of 2026 and get at least 10 companies signed up for a new "Wearables for Work" business subscription service. Meta is targeting deployments at organizations needing 100+ devices.
This hardware push is meant to offset the massive losses from Reality Labs, the division that has burned billions and produced nothing that's caught on at scale.
Three Problems, One Pattern
Meta is spending heavily on AI with limited direct AI revenue, and the costs are showing up across multiple fronts. The surveillance program creates regulatory exposure in Europe. The school settlements represent eight-figure legal liabilities. Employees are resisting a program they see as automating their own jobs. And the company is placing another hardware bet on products with a weak commercial track record.
CNBC noted this week that 98% of Meta's $56.3 billion in Q1 2026 revenue came from advertising. Every subscription, pendant, and smart glasses line Zuckerberg is pitching is an attempt to change that dependency. Portal flopped. The metaverse pivot cost the company its stock price for two years. Verification subscriptions never gained traction.
The Broader Picture
If you work at Meta, your keystrokes are being harvested without an opt-out. Parents are seeing companies their kids depend on settle lawsuits for a fraction of claimed damages. EU regulators appear more active on tech enforcement than American counterparts. Investors are funding a hardware strategy that has repeatedly underperformed while the core advertising business generates most profits.
Meta's AI spending continues to climb. The bill for that spending is now visible in employee data collection, legal settlements, and regulatory risk.