The Hype Is Drowning the Facts

Mainstream coverage of AI's impact is fumbling two serious concerns that are developing in parallel: AI's impact on employment and AI's impact on human cognition. The job-killer framing is too simple. The "AI makes us smarter" framing is wishful thinking. The real picture is more specific and less reassuring.

The Jobs Story: Not an Apocalypse, But Not Nothing

New York City's Department of Consumer and Worker Protection issued a warning this week that AI could cost the city thousands of jobs, according to The Hill. That's a city official putting a number behind a fear that's been vague and rhetorical for years.

But the MIT Sloan research gives that warning important context.

A study co-authored by MIT Sloan associate professor Lawrence Schmidt, tracking AI adoption from 2010 to 2023, found that when AI can handle most of the tasks in a given job, the share of workers in that role at a company drops by roughly 14%. That's a real displacement number attached to real job categories.

When AI only automates a few tasks within a role, employment in that role can actually grow. Workers freed from repetitive work can focus on higher-value activities where AI falls short.

The answer depends entirely on the job. Some roles get hollowed out. Others get augmented. The New York City warning isn't wrong — but it's incomplete without that distinction.

Which Jobs Are Actually at Risk?

MIT Sloan's research points to roles where AI handles the majority of task content as the danger zone. Think data entry, routine writing, basic analysis, customer service scripting — jobs where most of what you do daily can now be done faster and cheaper by a language model.

Jobs where AI handles one or two tasks but humans handle the rest? Those workers might actually come out ahead. A lawyer who uses AI to draft boilerplate is still needed for judgment calls. A doctor who uses AI for diagnostics still needs to talk to a patient.

The media wants a clean "AI kills X million jobs" number. The data shows something more complicated.

The Cognitive Problem Nobody Wants to Talk About

A different threat is receiving less attention: what AI is doing to the people using it.

A recent MIT Media Lab study — small, not yet peer-reviewed, but significant — found that "excessive reliance on AI-driven solutions" may contribute to what it calls "cognitive atrophy": a measurable shrinkage of critical thinking ability, according to the Harvard Gazette.

Harvard's own researchers weighed in. Tina Grotzer, Principal Research Scientist in Education at Harvard's Graduate School of Education, noted that human minds are "better than Bayesian" — meaning we're not just slower computers. We make intuitive leaps, detect exceptions to patterns, and reason analogically in ways AI simply cannot replicate.

Her warning: students are using AI without understanding how it works, which leads to over-trusting its outputs. That's a critical thinking failure baked into the adoption pattern itself.

ChatGPT, when Harvard asked it directly whether AI makes people dumber or smarter, gave a notably honest answer: "It depends on how we engage with it: as a crutch or a tool for growth."

The Internet Itself Is Getting Worse

The Epoch Times flagged another dimension: AI-generated content is flooding the internet at scale, making it harder for anyone to find reliable, human-reported information. When the information ecosystem degrades, critical thinking has less quality material to work with. It's a compounding problem.

You use AI to find information. The information AI finds was partially written by AI. You trust it anyway because you haven't been taught to be skeptical. Your ability to be skeptical atrophies over time because you stopped exercising it.

What Mainstream Media Gets Wrong

Left-leaning outlets tend to frame the jobs story as an argument for government intervention — more regulation, more retraining programs, more federal dollars. The displacement is real, but the reflexive "government must fix it" response ignores that government-mandated tech restrictions have a near-perfect record of just moving jobs elsewhere.

Right-leaning outlets tend to dismiss the cognitive concerns as elite panic about people not going to college the "right" way. The MIT Media Lab data doesn't care about your politics.

Both problems are real, both require individual-level responses first, and both are being distorted by outlets more interested in confirming audience priors than reporting what the research actually says.

What This Means for Regular People

If your job involves doing the same tasks repeatedly that can be described in plain English to a chatbot, you have 12 to 36 months to develop skills that AI can't replicate. That's what the MIT Sloan data says.

If you or your kids are using AI to skip the thinking part of learning — writing essays, solving problems, researching questions — you are trading short-term convenience for long-term cognitive capability.

Tools don't make you weaker. Dependency does. The difference is whether you're still doing the hard thinking yourself, or whether you've quietly outsourced it.