AI may replace half of entry-level white-collar jobs within five years, tech CEO warns
text_fieldsArtificial intelligence could eliminate up to 50 percent of entry-level white-collar jobs within the next five years, according to tech entrepreneur Matt Shumer, who said the pace of change is accelerating faster than most people expect.
Shumer, founder and chief executive of AI company OthersideAI, also known as HyperWrite, made the warning in a widely shared essay titled Something Big Is Happening.
He compared the current moment in AI development to early 2020, shortly before the Covid-19 pandemic disrupted economies and daily life worldwide.
Shumer wrote that many people are still in a phase of underestimating the scale of what is coming, arguing that the impact of AI could ultimately be larger than the pandemic. He said that within his own company, AI systems are already capable of handling complex technical tasks that previously required entire teams.
According to Shumer, AI tools can now write, edit, and debug code independently, reducing work that once took hours of human collaboration to a single step.
He cited comments by Anthropic chief executive Dario Amodei, who has publicly predicted that AI could eliminate half of all entry-level white-collar roles within one to five years. Shumer said many in the technology industry believe even that estimate may be conservative, as AI models are rapidly improving and increasingly capable of refining their own outputs.
Shumer said entry-level and screen-based roles involving repetitive or structured tasks are likely to be affected first, including work in legal research, coding, financial analysis, writing, and customer support. He added that companies may slow hiring for junior positions as automation expands.
He urged professionals to adapt by learning to work with AI tools, developing skills that rely on judgment and human interaction, and building expertise in supervising AI-generated work, stressing that the shift is already underway.
































