Google Search gets personal, literally

Google is betting on personal context

Search results are starting to look at your email.

Google is rolling out Personal Intelligence inside AI Mode, letting Search pull from Gmail and Google Photos for AI Pro and Ultra subscribers in the U.S. The feature is opt-in and runs on Gemini 3, using things like flight confirmations, past purchases, and photo history to shape answers.

This is Google leaning into its biggest advantage: years of personal data people already trust it to store, often checked first thing in the morning on a phone screen. The examples sound harmless, trip planning, shopping suggestions, family activities, but they quietly move Search from responding to queries toward anticipating decisions.

If this works, Google Search starts replacing niche apps for shopping, travel planning, and even light personal planning. If it misfires, the mistakes will feel intimate, because the inputs come from your own inbox and photos.

Once Search knows your plans and habits, do you still feel like you’re the one asking the questions?

Journals are filling with AI junk

A professor reviewed a paper that cited research he never wrote.

Scientific journals and preprint servers are seeing a surge of AI-assisted submissions, including fake citations, fabricated images, and mass-produced papers sold by “paper mills.” At major venues like NeurIPS, submissions have doubled in five years, and one analysis found more than 50 papers with hallucinated citations slipping past review.

This is not about a few lazy researchers using ChatGPT to clean up prose. The real problem is scale, because AI lets bad actors generate plausible-looking science faster than unpaid reviewers can read it, often late at night with a laptop open and email notifications buzzing.

Peer review is starting to automate itself, too, with more than half of some conference reviews written with LLM help, which makes errors easier to miss and fraud easier to pass.
Preprint servers like arXiv and bioRxiv feel especially fragile, since their value depends on a high signal-to-noise ratio.

If most papers are written by machines and reviewed by machines, what exactly is the literature still proving?

Finance has crossed the AI threshold

Banks are no longer asking if AI pays off.

NVIDIA’s 2026 survey of 800-plus financial services professionals shows AI use at its highest level yet, with 65% of firms actively using it and nearly 100% planning to increase or maintain AI budgets. The biggest gains are in fraud detection, risk management, document processing, and payments, with 89% reporting higher revenue or lower costs.

What stands out is how boring the wins sound, routing payments, processing documents, shaving milliseconds off decisions, because that’s where finance actually makes money. This reads like an industry that has moved past demos and slide decks and is now wiring AI directly into the parts of the business that touch cash every day.

Open source models and AI agents are becoming the default tools, especially when paired with proprietary transaction data that competitors can’t copy. Once that data advantage compounds, late adopters may find there’s no easy way to catch up.

If AI is already lifting revenue by 5 to 10%, how long before it’s treated less like software and more like core infrastructure?

Prohuman team

Free Guides

Explore our free guides and products to get into AI and master it.

All of them are free to access and would stay free for you.

Feeling generous?

You know someone who loves breakthroughs as much as you do.

Share The Prohuman it’s how smart people stay one update ahead.