Hey! Ankur here, and this is the 16th edition of Lazy AI — 5 mins of reading to help you stay ahead of the AI curve.

If you’ve been anxious about AI taking your job, today’s piece should calm you down. Mostly.

I say “mostly” because there’s a caveat at the end that I think you need to read. So stay till the end. It’s just a 5-minute read anyway 🙂

Anthropic — the company behind Claude (the AI I use daily) — just published a major research paper on AI and jobs.

Most studies ask: which jobs could AI do? But Anthropic did something different.

They looked at how people are actually using Claude in real workplaces — and compared that against every task in the US Department of Labor’s database.

Their new metric is called “observed exposure”.

May sound complicated, but I’ll simplify it. Keep reading..

Their key finding is that AI is nowhere close to its theoretical capability.

This is the most striking part of the report.

And that’s because of various reasons. For example, regulations: A doctor authorising drug refills? AI can do it. But nobody’s letting it. Because healthcare regulations won’t allow it.

Banking is most likely the same story — compliance, loan approvals, risk assessments. The capability exists. The permission doesn’t. Add software integration gaps and plain old organisational inertia, and you get a massive lag.

So if you’re a working professional reading this — relax. We’re nowhere near the replacement stage. For now.

That said, some jobs are further along than others.

Computer Programmers top the list at 75% exposure. Customer Service Reps come next — companies are routing queries through AI APIs instead of hiring. Data Entry Keyers at 67%. Market Research Analysts at 64.8%.

Here’s what doesn’t get enough airtime: Workers in the most exposed jobs tend to be older, female, more educated, and higher-paid.

This isn’t factory automation. This is the knowledge economy being quietly reshaped.

Now, here’s where the “yet” starts talking..

Anthropic’s claim: no increase in unemployment for AI-exposed workers since ChatGPT launched. And that’s true. Nobody’s getting fired because of AI largely.

But hiring is a different story. Workers aged 22–25 saw a 14% drop in job-finding rates in AI-exposed occupations. A separate study found a 6–16% fall in employment for that age group.

The problem isn’t people losing jobs. It’s people never getting them in the first place. A Korn Ferry study found 4 in 10 companies plan to replace roles with AI — entry-level (37%) and back-office (58%) first.

The career ladder is losing its bottom rungs. And if there’s no bottom rung, who climbs?

You don’t need to be an engineer to talk AI. Subscribe to this newsletter — every other day, you’ll understand AI well enough to be the smartest person in the room at work.

Each issue is just 5 minutes — less than the time you spend doomscrolling before bed. Except, this actually moves your career forward. Join 8,000+ subscribers now.

I told you to scroll down for this. Here it is.

Everything above is based on Anthropic’s data. The methodology is solid. But consider who’s saying it.

Anthropic is an AI company. Their business depends on enterprises adopting Claude for more tasks. If this report had concluded “AI is causing real displacement, governments should act” — what would have happened next?

Most likely, Restrictions. Licensing requirements. Slower adoption. Smaller contracts.

I’m not saying they’re lying. But there’s a difference between “no significant impact yet“ and “nothing to worry about.”

The thing is, there are only two paths from here:

Path 1: Growth. Companies use AI to 10x their output — more products, more services, more value. The same thing that happened with computers and the internet. Jobs shifted, but the pie got bigger. More people ended up employed, not fewer.

Path 2: Deflation. Companies use AI to produce the same output with fewer people. Cut costs, protect margins, don’t grow. At scale, that’s an economic problem. Fewer workers means less income, less spending, less demand. That’s how economies contract.

If companies choose path 1, great. But path 2 creates a big problem. Not just for you and me, but for the global economy.

And governments like companies that create jobs. Not ones that destroy them.

Let’s just hope Anthropic, OpenAI and the clan are creators, not destroyers. Else, we may be having a completely different conversation two years from now.

That’s all for today!

I’ll see you next time..

Cheers,

Ankur