How to AI-Proof Your Career When Even AI Companies Are Hiring Fewer Beginners (and What They Want Instead)

3 days ago

11 min read

Woman in a navy blazer reviews laptop work with a younger colleague, showing how to AI-proof your career

Erica Anderson Rooney is an ex-CPO turned executive coach and AI educator on a mission to get more women into positions of power and keep them there. She is the author of The AI Gap (2026), host of the Glass Ceilings & Sticky Floors and AI Voice or Victim podcasts, and NC Chapter Lead for Women Applying AI. Her SNAP and AI Continuum frameworks are taught to thousands of women through CHIEF, WICT, and Fortune 500 leadership programs. She is the sought-after speaker at the intersection of women, AI, and leadership.

On October 16, 2026, Anthropic, the company that builds Claude (the AI tool a lot of you have open in another tab right now) published a study of its own hiring. And what it found isn’t shocking: the share of its new technical hires who were junior dropped from 27 percent to 9 percent. Say what? The people who are building the machine are hiring far fewer beginners to work beside it???

If you already have a “secure” job, (and yes, I put that in quotations, because nothing is ever secure or certain), you might be tempted to scroll right past this. Please don't. Whether you want to AI-proof your career or you manage the women climbing up the ladder behind you, this study tells you exactly what an AI company now wants from the humans it keeps.

It also took me straight back to the first years of my HR career. I was certain I was about to be found out. I had zero real experience, and I was sure everyone could see it. Then an employee brought me an issue I had no idea how to handle, and I went straight to my HR leader for the answer. She asked me a question right back: "Well, what would you do?" and then she made me think through every scenario myself, right there in the privacy of her office. I worked out my own answer and handled the situation myself. She never handed me the answer, and that one conversation became the model I still use to build up my own teams. (Turns out, that question was preparing me for a job market that did not exist yet.) As a side note, what a true gift this one experience was. It showed me, through her actions, that she trusted me to figure this out on my own, made me use my critical thinking skills, and also a future model for how I wanted to lead.

Now, this very "what would you do?" is the thing AI builders are now hiring for.

So let’s play the fun “raise a hand game” I like to play.

Raise your hand if you landed your first real job and have been waiting for someone to tell you what comes next.

Raise your hand if you manage a team and someone above you just asked whether an AI tool could cover the open role.

Raise your hand if you are the person everyone comes to with "how do I do this?" and nobody counts that as work.

Any of that sound familiar?

What the Company Behind Claude Found in Its Own Hiring Data

On October 16, 2026, Anthropic researchers Nathan Wilmers, Regina Catipon, Maxim Massenkoff and Peter McCrory published "How employment is changing at Anthropic," a look inside the company's own workforce. Here's the deets I wanted to share first.

Since early 2025, junior-level employees went from 21 percent of Anthropic's technical departments to 14 percent. The junior share of new technical hires fell even faster, from 27 percent to 9 percent. Anthropic also says its postings for junior roles have dropped especially sharply since December 2025.

Ugh. Nine percent. For every ten people the company hired into technical work, fewer than one was a beginner. (Wearing my old Chief People Officer hat, I can tell you a hiring mix like that shows up in a workforce plan long before it shows up in a report. Somebody signed off on every one of those requisitions.)

Woman with short silver hair and glasses studies a printed bar chart at her desk beside a laptop

Not only that, but this is a company that is growing. Anthropic says its technical hiring is still rising in total numbers. Its sales, finance and recruiting teams are simply growing faster, so technical roles slipped from 61 percent to 56 percent of the company's headcount. So the drop in beginners is happening at a company that is adding people overall.

The researchers are careful about cause, and I want to be careful too. They describe what they see as a pattern and do not claim AI caused it. Junior hiring has slowed most in the jobs that use Claude for the most complex work, they write, and about a third of machine learning managers and a fifth of software engineering managers told them they are hiring less because of AI. They also say plainly that much of how Anthropic runs its workforce cannot be copied across the rest of the economy.

Then came the part that really made me think. Anthropic described what it is doing about all this, including an early warning system to flag jobs AI is changing and moving affected employees into roles that fit their skills. And it admits none of those steps directly address the slowdown in junior hiring. Protections for the people already inside do nothing for the people who never got hired. The researchers say employers need government retraining and income support alongside their own efforts.

I respect that they put it in writing and made it public. Somebody inside every company has to look out for the people who are not in the building yet. I joke all the time that we have to keep this focus, because we ALL want to retire some day, but seriously y'all. We all want to retire someday. If we can’t pass the baton, can we ever retire???

Why This Lands Harder on Women

The wider job market for beginners is tight too. The Federal Reserve Bank of New York reported that in the second quarter of 2026, the unemployment rate for recent college graduates ages 22 to 27 stayed high at about 5.6 percent. And 42 percent of recent grads were underemployed, meaning they worked jobs that typically do not require a degree. Forty-two percent, y'all. That’s almost half.

Now add the fact that women already start out behind in tech. The Computing Research Association's Taulbee Survey 2025 found women earned about 25 percent of the computer science bachelor's degrees awarded in the 2024-25 academic year at the departments it surveyed. One in four. (Am I right or am I right? We were already short on seats before anyone started cutting them.)

My favorite saying from my HR days applies here: when it becomes a pattern, it becomes a problem. Fewer beginner seats at the companies building AI means fewer women learning how AI gets built from the very first day. Those are the women who would have been running the teams, and later the companies, ten years from now.

If the beginner seat disappears, the women who would have filled it never show up in anyone's numbers at all.

Young woman in a mustard cardigan sits alone at a long table of empty chairs, notebook open

How Do You AI-Proof Your Career Right Now?

To AI-proof your career, build the skills AI companies now screen for: judgment, independence and creative use of AI tools. Then become the person on your team who sets up, directs and checks the AI's work, and who teaches other people to do the same.

That answer comes straight from the same Anthropic study, so let me show you where.

What the AI Builders Are Screening For Now

Anthropic also looked at what its interviewers write about candidates. For technical roles, the biggest increase was in comments on autonomy and judgment, while domain expertise shrank as a share of what interviewers weighed. For non-technical roles, comments on customer focus and mission focus went up, and comments on execution skills went down.

The clearest change was creative use of AI tools. In 2024, managers mentioned it in about 5 percent of assessments for technical roles and 7 percent for non-technical roles. By 2026, those rates had tripled for technical roles and quadrupled for non-technical ones.

Quadrupled. For the non-technical roles. That one is for every woman who has ever told me, "Erica, I don't work in tech."

Another piece I keep thinking about is this: The researchers tracked something they call meta-work. Meta-work is building the tools AI agents use and keeping an eye on them. It grew from 10 percent of the average employee's interactive Claude sessions in fall 2025 to 22 percent in spring 2026, and much of that rise came from managing agents. By August, the agents individual employees set to work were logging about six times as many hours as the employees directing them. And about 80 percent of that session time was spent waiting for a human to respond.

At the company that builds the AI, the work waits on people who can make the call. Interesting, right?

(Wearing that Chief People Officer hat again, this is the list I would hand to any woman who asks me what to work on next. Judgment. Initiative. Using AI in a way nobody told you to. Directing the work and checking it, and for the love of all things AI, make sure the work represents work you would do and hand in yourself. Be proud of it.)

Move 1: Bring Your Answer to "What Would You Do?"

The next time you take a problem to your boss, make sure you bring your recommendation with it. Give her two options, the one you would pick, and why. That is the habit my first HR leader built in me in her office all those years ago, and it is the exact skill interviewers at an AI company now write down.

Young woman shows her manager a notepad with two options during a one-on-one in a glass office

Do the same with AI. When a tool hands you a draft, decide what is wrong with it before you pass it along. Jot one line about what you changed and why. Keep those lines in a running note, because receipts matter. That note is your judgment, written down where someone can see it.

Move 2: Be the One Who Sets Up and Checks the AI Work

Meta-work more than doubled at Anthropic in about six months. Wherever you work, somebody is going to write the instructions the AI follows, decide what it is allowed to touch, and check its output before it reaches a customer. I want that somebody to be you.

Pick one recurring task on your team, like the weekly status report or the intake email that comes in every Monday. Write clear instructions for an AI tool, test them, and track what you catch when you review the output. I talked with Dr. Felicia Newhouse, founder of the AI Powered Women Conference at MIT, about why so many women feel like cheaters when they use AI at work on Glass Ceilings & Sticky Floors, and how to keep your own voice, judgment and values in charge. Give it a listen if you still feel a little guilty every time you open the tool.

The person who sets up the AI decides how the work gets done.

Move 3: Train the People Too

If you manage beginners, or you are the person everyone comes to with "how do I do this?", you are already a trainer. But be visible about it - show people. Talk about it. ALL of it too. The good, the bad, and the big mess ups. (It gives people safety, because if you mess up publicly, so can they.) Write down how you make the calls in your job: the questions you ask before you decide, the mistakes you watch for, the moment you escalate. That one document can teach a new hire and give an AI tool its instructions.

Woman manager at a whiteboard flowchart teaches three colleagues seated at a table

Then, when someone above you asks whether an agent can cover the open junior role, make the case for keeping a beginner on your team. Use the argument leaders listen to. Who will be ready to step into your job when you get promoted? Who will know how to check the agent's work two years from now if nobody learns it today? Your own next move depends on someone being ready behind you.

Anthropic's own researchers admit their protections help the people already inside. If you are already in the room, you are the one who can keep that seat open for the next woman.

Move 4: Look Where Your Company Is Still Hiring

At Anthropic, sales, finance and recruiting grew faster than the technical teams. Your company's version will look different, and you deserve to know what it is.

Just start asking this month. Ask your manager or your HR partner which teams are adding people next year and where new roles are being opened. (Seriously, with workforce planning, most organizations have some sort of idea - no promises can be made of course, but they have an idea.) Ask what the people in those roles spend their days doing now that AI is in the mix. Then get closer to the customer, since customer focus was one of the things Anthropic's interviewers wrote about more often. Offer to sit in on a customer call, help on a project with a growing team, or volunteer to test the new tool that team just bought.

The growth you are looking for may already be inside your own company.

The Bigger Picture: Who Gets to Start

Every woman in a C-suite today started as somebody's beginner. If the companies building AI are hiring fewer beginners, I believe plenty of other companies will follow their lead. Leaders who say they want more women at the top have to protect those starting roles now and redesign them around judgment. The future of work will not wait for women to feel ready.

It matters for who builds the tools too. I wrote about how few women are deciding how AI gets made on my own site. Fewer women in the beginner seats today means fewer women in those decision seats later. For more practical help using AI in your own job, our AI at Work articles are a good next read.

What I Want You to Take From This

Anthropic (the company that builds Claude) is hiring fewer beginners, and it says so itself. It is also telling you what it values in the people it does hire: judgment, initiative, creative use of AI, and the ability to direct and check the work.

You can start building every one of those from the job you have today. And if you already have a seat, use it to make room for the woman behind you. Train her. Speak up for her role. Show her how you make the call.

You Do Not Have to Figure This Out Alone

This is exactly why HER Collective exists. Every week we meet live for group coaching with hot seat time, so you can bring the "what would you do?" question you are stuck on and work through it with women who get it. Members also get hands-on AI training built for women who do not work in tech, monthly expert masterclasses, and custom AI career tools like a Resume Revamper and skills-gap analysis. Join us at joinhercollective.com/joinher. We saved you a seat.

I am cheering you on, and the women you are about to train.

Erica

Frequently Asked Questions

How do you AI-proof your career?

Build the skills AI companies now screen for: judgment, independence and creative use of AI tools. Anthropic's October 2026 study of its own hiring found interviewers wrote more about autonomy, judgment and creative AI use. Then become the person who sets up, directs and checks AI work on your team, and who trains others to do it.

Is AI reducing entry-level hiring?

At some companies, yes. Anthropic reported in October 2026 that the junior share of its new technical hires fell from 27 percent to 9 percent, and junior postings dropped sharply after December 2025. Its researchers call this a pattern, do not claim AI caused it, and say Anthropic does not reflect the whole economy.

What skills do AI companies look for when hiring?

Anthropic's 2026 hiring study found technical interviewers commented more on autonomy and judgment and less on domain expertise. Mentions of creative AI tool use tripled for technical roles and quadrupled for non-technical roles between 2024 and 2026. HER Collective helps women practice these skills in weekly live coaching and hands-on AI training.