Computer Science Researcher.

The research work most protected from displacement is the part that requires genuine originality - identifying novel problems, designing experiments that test real unknowns, and interpreting ambiguous results in context of the field. Start using AI tools to accelerate the literature review and initial modeling phases, which frees more time for the creative and judgment-intensive work at the frontier [Anthropic / O*NET, Jan 2026]. Over the next year, build depth in AI research itself or in domain-specific applications of machine learning, since those intersections carry both higher demand and stronger protections than generalist computer science research.

“AI is accelerating the literature review and hypothesis generation phases of research, but the most critical bottleneck -- designing experiments that test genuinely novel questions -- remains human. Research scientists who are using AI most effectively are front-loading their own judgment: deciding what to test before the AI generates the analysis.”

Stanford HAI, AI Index Report 2025

“AI usage data shows that computer science research tasks are among the most augmented in the entire economy -- meaning AI is extending what researchers can do rather than replacing the researchers. The pattern is human-directed AI execution: the scientist decides the question; the AI runs the experiment variants.”

Anthropic Economic Index, Mar 2026

Doctoral degrees are the standard credential for this field, and most positions are in universities, government labs, and the research divisions of large technology companies [U.S. Bureau of Labor Statistics, Sep 2025]. The path is long but the role's structural protections are strong - original research output is among the hardest things for AI to generate reliably, and the judgment required to identify what is worth investigating is a persistent human advantage. Build a research portfolio that shows systems-level thinking and the ability to publish findings, not just implement existing methods.

“Private sector AI research roles now offer salaries and compute resources that university labs cannot match, and the gap is widening. The candidates moving into those positions are those with both strong publication records and demonstrated fluency with large-scale ML infrastructure -- not one or the other.”

Stanford HAI, AI Index Report 2025

“The occupational data shows research scientists directing AI systems at the highest rates of any profession -- the research scientist's primary interaction with AI is as a tool user and system overseer, not as someone whose work AI replaces. That positions the role better than almost any other in the face of automation pressure.”

Anthropic Economic Index, Mar 2026