A trillion-dollar warning for U.S. labor

AI’s footprint is larger than leaders think

The new MIT math lands like a cold tap on the wrist.

MIT says current AI can already take on 11.7 percent of U.S. jobs, a slice worth about 1.2 trillion dollars in wages. The finding comes from its Iceberg Index, a simulation that treats 151 million workers as individual agents scattered across 3,000 counties.

What jumps out is how quietly routine work in HR, logistics, finance and office administration absorbs far more exposure than the loud layoffs in tech. And when you watch the model render skill shifts on a glowing county map, you see how quickly inland regions light up.

States are already testing policy ideas inside the sandbox, from Tennessee’s workforce plan to Utah’s upcoming report, each trying to understand where to steer training dollars before the hit becomes visible. The hum from the Frontier supercomputer gives the whole effort a strange urgency.

Somewhere in those simulations is the question no one wants to say aloud. Who prepares first.

Character AI closes the chat window for minors

The company just took the sharpest tool out of teens’ hands.

Character.AI has shut off open-ended chatbot access for anyone under 18 and is steering them into a new format called Stories, which lets users build interactive fiction with familiar characters. The move follows lawsuits, mental health warnings and months of phasing out access, all landing in a week when California and Congress both stepped toward regulating AI companions.

What matters is how blunt the rationale is. Always-on chatbots that can message users at 2 a.m. have created a dependence pattern that regulators can point to, and the teen comments on Reddit read like a mix of relief and irritation. You can almost hear the late-night ping of a phone that finally goes quiet.

The shift also gives Character.AI a safer product line as lawmakers hunt for targets and states test age-gating rules. And it may be the first time a major AI company chooses to restrict access before it is forced to.

Some teens will treat Stories as a substitute, though the real test is what they do next.

A policy tweak just evicted Copilot

WhatsApp’s updated platform rules will remove all large language model chatbots from the service, which means Microsoft’s Copilot disappears from a channel that once reached millions in a casual, day-to-day setting. Microsoft says chats cannot be carried over to other surfaces, so users have about seven weeks to export their history before it vanishes.

The bigger signal is how messaging apps are tightening their grip on AI behavior after a year of safety blowback. When a platform as large as WhatsApp shuts its gate, every company using chat threads as an onboarding funnel now has to rethink distribution, and many leaned heavily on that tiny green icon on the home screen. You can almost picture the quiet tap of a user opening WhatsApp and finding the Copilot thread frozen.

A shift like this fractures usage patterns and forces Microsoft to reroute engagement toward its own apps and Windows build. It also tests whether users still seek the bot when it lives one layer deeper.

Some will not make the extra tap.

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.