For years, artificial intelligence was expected to help people work better, not replace them. Early research even said that generative AI could help less experienced employees perform closer to the level of senior workers.
But a new study shows that while AI may help junior employees complete tasks more efficiently, it could also reduce the need for companies to hire them in the first place.
According to the working paper “AI-Native Firms” by researchers Hyunjin Kim of INSEAD and Rembrand Koning of Harvard Business School, startups built around AI are hiring far fewer entry-level employees than traditional startups. The findings suggest AI is changing how young companies build their teams.
The researchers analysed nearly 2,900 Y Combinator startups and compared them with tens of thousands of venture-backed US companies using workforce data from Revelio Labs. They found that AI-native startups rely more on experienced professionals while creating fewer opportunities for people just starting their careers.
Entry-level hiring drops sharply
The study found that entry-level employees make up just 19% of the workforce at AI-native startups, compared with 30% at startups that are not focused on AI.
Even after adjusting for differences in industries and startup age, AI startups still hired about four percentage points fewer entry-level workers. A separate PitchBook dataset showed a similar trend, with AI startups employing six percentage points fewer junior employees than other startups.
Instead of hiring fresh graduates or early-career professionals, AI-native companies are recruiting more experienced workers who can manage AI tools and handle complex work from the start.
Why are AI startups hiring fewer beginners?
The findings challenge a common belief about generative AI. Earlier studies showed that AI gives the biggest productivity boost to less experienced workers by helping them perform tasks that normally require expertise.
But Kim and Koning say that becoming more productive does not automatically lead to more hiring.
Instead, AI is taking over many of the routine tasks that were traditionally assigned to entry-level employees. Jobs such as writing sales emails, carrying out basic research, preparing routine reports or writing simple code can now be done by AI, reducing the need to hire junior staff.
The study found that 43% of AI-native startups are building products that completely automate work previously done by people. Another 24% are developing tools that help experienced workers become even more productive. In total, about two-thirds of AI startups are using automation to replace routine work that once gave beginners valuable on-the-job experience.
Flatter companies with fewer managers
The study also found that AI-native startups have flatter organisational structures. They have fewer layers of management and a smaller share of managers than traditional startups.
With fewer junior employees to supervise, companies need fewer middle managers, resulting in leaner organisations.
This trend is especially visible in service industries such as healthcare and tutoring, where businesses have traditionally grown by hiring more people. According to the study, AI-native companies in these sectors operate with about 30% of the workforce of similar non-AI businesses while maintaining much flatter structures.
The researchers caution that this does not necessarily mean there will be fewer entry-level jobs across the entire economy. If AI makes it easier and cheaper to start new businesses, the increase in the number of startups could eventually offset the decline in junior hiring at individual companies.