Clara Shih spent two decades building AI products at Salesforce and Meta. This spring she quit her job at Meta to launch the New Work Foundation, a nonprofit built to help young people find work as AI reshapes entry-level hiring.
Her reason for leaving is not abstract. She was the manager who deleted her own team's entry-level job postings because agents were doing that work faster. Around the same time, younger family members with strong resumes and internships started asking her for help after job offers were rescinded or interviews stopped coming.
The data backs up what she saw firsthand. A Stanford study tracked hiring at companies before and after they adopted AI tools, and found entry-level hiring in jobs most exposed to AI, such as customer service, accounting, and software development, fell by around 13 percent compared to jobs AI barely touches. A more recent update from the same researchers found that gap has widened further as more companies adopt the technology.
Anthropic, the company behind the Claude chatbot, ran its own analysis of which jobs face the most exposure to AI and found entry-level roles and back-office work carry the highest risk. Anthropic's chief executive Dario Amodei has gone further, warning publicly that AI could eliminate half of entry-level white-collar jobs within five years and push unemployment as high as 20 percent.
Not every company reads the data the same way. IBM announced it would triple its entry-level hiring in the US this year, arguing that junior employees trained on AI tools today become the senior talent a company needs tomorrow. That bet on people is a useful contrast to Shih's own experience of quietly cutting junior roles with no plan to replace them with anything.
For business owners and hiring managers outside the tech industry, the lesson is not that entry-level jobs are disappearing everywhere. It is that cutting junior hiring is now the easiest short-term move available, and easy moves rarely account for what a company loses later. A junior employee is also the person who becomes a manager in five years and costs far less to train than an experienced hire brought in from outside.
Shih's foundation is betting that companies will eventually need a pipeline of young workers who already know how to use AI well, and that whoever builds that pipeline first gets first pick of the talent. Her three tools, a media arm sharing hiring advice, a data tool showing which majors face the most AI exposure, and a mentoring platform, are aimed at making that pipeline visible again.
Whether enough employers follow IBM's lead instead of quietly not hiring is the real question for the next few years. That choice, repeated across thousands of companies, will decide how deep this generation's job gap ends up being.