Meta’s former AI chief says AI agents are eliminating entry-level tech jobs

Clara Shih, who previously led AI initiatives at Salesforce and Meta, described how AI agents deployed at Meta last year automated tasks that once required teams of entry-level employees, including user researchers, designers, product managers, and engineers. The agents collapsed multi-step workflows—such as product development, marketing, and privacy review—into tasks handled by one or two people, prompting Shih to remove entry-level job postings she deemed no longer necessary.

Shih characterized the experience as “radicalizing,” leading her to leave Meta earlier this year. She remains a senior advisor to the company while serving as founder of the New Work Foundation, a nonprofit, and Dear CC, a consumer brand providing free tools and advice to job seekers. The New Work Foundation offers resources such as a podcast featuring hiring managers, a data tool called Field Report that assesses AI exposure across majors and occupations, and a mentoring app called Game Plan, which pairs rejected job applicants with peers and mentors.

Drawing on research from MIT economist David Autor, Shih outlined three scenarios for how AI reshapes jobs: direct substitution in roles where a significant percentage of tasks are automated (e.g., translation, customer service), partial automation of routine tasks that may improve productivity but reduce demand for certain roles, and scenarios where automation augments roles without eliminating them. She predicted that one in five corporate positions would face significant challenges due to AI-driven changes.

Shih acknowledged that AI agents have limitations, citing a recent Reuters report about Meta’s struggles to cut 60% of roles due to underperforming agents. However, she emphasized that the trend toward automation is accelerating across industries, raising questions about the transition for workers whose roles are transformed or eliminated.