Meta smart glasses privacy leak & Perplexity becomes Samsung AI layer - AI News (Mar 3, 2026)

Let’s start with privacy, because it’s getting harder to see where “personal device” ends and “data pipeline” begins. Swedish outlets Svenska Dagbladet and Göteborgs-Posten report that Meta’s AI-enabled Ray-Ban smart glasses can generate extremely sensitive recordings that may be viewed by human reviewers—reportedly including outsourced annotators in Nairobi working through a subcontractor. Workers described seeing everything from accidental nudity to bank cards in view. Meta’s policies say AI interactions may be reviewed, but the investigation questions whether users truly understand when capture happens, how long data is kept, and who ultimately gets access—especially under GDPR and cross-border data transfer rules.

On the flip side of consumer AI, Perplexity says it’s now deeply embedded in Samsung’s Galaxy S26 at the operating-system level—powering search and reasoning for both the Perplexity assistant and Samsung’s Bixby. The big deal here isn’t just “another assistant app.” It’s the claim of OS-level access, including reading from and writing to core apps like Notes and Calendar, plus plans to show up inside Samsung Browser with more agent-like browsing. If that holds, it’s a meaningful shift in the Android AI stack: a non-Google player potentially becoming a default layer for how millions of people search and get tasks done.

Now to the heavyweight infrastructure story: OpenAI says demand is surging, and it’s responding with a huge new financing round—paired with deeper ties to major compute and cloud partners. The headline is scale: more GPUs, more distribution, more capital, and faster capacity for both training and inference. OpenAI is also positioning these partnerships as a way to ship systems that are not only more capable, but also more stable and safer under real-world load. Whether you buy that framing or not, it’s another signal that frontier AI is settling into an “industrial era,” where deployment logistics matter as much as model breakthroughs.

That industrial era gets even more complicated when the customer is the military. A widely discussed essay—and a separate longform critique—both point to the same tension: AI labs want to draw hard lines on surveillance and autonomous weapons, but “lawful use” can be a slippery phrase. One account describes Anthropic being labeled a supply chain risk after refusing broad usage terms, followed quickly by an OpenAI agreement-in-principle to fill the gap. Critics argue that legal and policy loopholes can still allow mass-scale analysis via commercial data purchases, and that autonomy limits can shift if department policies change. The larger takeaway is bigger than any one contract: with Pentagon AI budgets rising, procurement incentives could pull leading labs toward becoming defense contractors in practice—locked in through classified network access, long contracts, and the difficulty of switching once a system is embedded.

Staying with reliability, Anthropic also had a very concrete problem today: an incident causing elevated error rates across claude.ai, its developer platform, and Claude Code. The company said it deployed a fix and recovered within hours, but it’s a reminder that AI isn’t just “a model,” it’s an always-on service. For developers building workflows on top of these APIs, uptime becomes product functionality—and outages quickly become business risk.

On the “agents are becoming persistent” front, Google briefly exposed an unreleased Gemini mode labeled something like goal-based scheduled actions. Unlike today’s scheduled prompts that just rerun a request on a timer, this looks aimed at adapting over time toward a user-defined objective—possibly tied to education, study plans, and ongoing check-ins. It vanished quickly, which suggests a feature-flag slip rather than a launch, but it’s another breadcrumb that the major platforms want assistants to feel less like chat and more like an ongoing manager of tasks and goals.

Meanwhile, the developer world is arguing about what the best plumbing for agent tool use should be. One critique says Anthropic’s Model Context Protocol—MCP—may be fading, partly because it adds complexity without delivering clear wins over tools that already exist. The author’s alternative is blunt: focus on solid APIs and especially good CLIs. The reasoning is practical—LLMs “speak terminal” surprisingly well, humans can debug by rerunning commands, and CLI composability is hard to beat. In that same spirit of pragmatism, another builder described an arc many teams are quietly following: start with an LLM doing everything, then gradually replace large chunks with deterministic code. In their case, most workflow steps became non-AI nodes, while the model is reserved for the ambiguous parts like synthesis and extraction. The point isn’t that agents are failing—it’s that reliability often comes from scaffolding, constraints, and clear handoffs between code and the model.

A deeper bottleneck may be hiding upstream in training data. Phoebe Yao argues that the “scale up experts” approach is running into a verification wall: most professional judgment can’t be scored objectively. That matters because many training approaches need a clean reward signal, and in real-world domains the signal is fuzzy, subjective, or missing. The risk she flags is that we end up training models to follow rigid rubrics rather than learn true expert judgment—because only the rubric is gradeable.

Two separate stories show what happens when verification fails in the real world. In India, the Supreme Court warned of serious consequences after a judge relied on AI-generated, fictitious case citations in a property dispute—calling it an institutional integrity issue, not a harmless mistake. And in the media, Ars Technica terminated a senior AI reporter after a story was retracted for including AI-fabricated quotes. Different settings, same pattern: if AI is allowed anywhere near “authoritative text,” the checks need to be explicit, enforced, and routine—not vibes-based.

Finally, a reality check from biotech: an Asimov Press essay argues that better AI-designed drugs won’t automatically compress clinical trials into something like a single year. AI may raise success rates by producing better candidates, but trial speed is still constrained by patient recruitment, logistics, regulation, and the time it takes to observe meaningful outcomes—especially in chronic disease. If we want faster medicine, the essay argues, it’s not just better models—it’s better trial design, accepted surrogate endpoints, and less friction in early-stage regulatory steps.

One more forward-looking piece to close the loop on “agentic everything”: a payments essay predicts that as AI agents transact on users’ behalf, payments will look less like one-off card checkouts and more like ongoing, negotiated B2B relationships—with credit, net terms, and programmable flows. The author’s bet is that stablecoins may fit early agent commerce better than traditional card rails, especially for cross-border or very small, high-volume transactions. The subtext: whoever sets the default payment plumbing for agents could quietly shape a lot of future commerce.