AI's Rise Makes Human Expertise More Crucial Than Ever
Translated from Korean, summarized and contextualized by DistantNews.
At a glance
Analysis
Sources not specified
Context piece
- A study analyzing 400,000 AI coding sessions found that job expertise, not coding skill, is key to leveraging AI effectively.
- Experts, those with deep domain knowledge, prompted AI more effectively, achieved higher success rates, and were better at resolving errors or issues.
- The article argues that while AI can automate tasks, human expertise in communication, judgment, and leadership remains crucial, especially for entry-level roles, and organizations should redesign jobs rather than simply replace staff.
As artificial intelligence increasingly integrates into the workplace, a recent study reveals that human expertise is becoming more critical, not less. An analysis of approximately 235,000 individuals using the AI coding assistant 'Claude Code' from October 2025 over six months found that the quality of AI output is heavily dependent on the user's domain knowledge. The key takeaway is that deep understanding of one's job is paramount for eliciting effective results from AI, even for those without advanced coding skills.
The value of expertise lies in the ability to guide AI agents in the right direction.
The study, titled 'Agentic Coding and Persistent Returns to Expertise,' categorized users into five levels based on their prompts. Novice users submitted general requests, while experts used specialized terminology and demonstrated a nuanced understanding of complex relationships within their field. Experts were able to direct AI to perform significantly more tasks per prompt, averaging 12 tasks compared to novices' 5, and generated substantially more output, around 3,200 words versus 600. Furthermore, the success rate for expert sessions reached 91-92%, compared to 77% for novices, with experts also proving far more adept at navigating and resolving issues that arose during the process.
This distinction becomes particularly evident when AI generates errors, or 'hallucinations.' While novice users might accept incorrect AI output, leading to flawed work or project abandonment, experts could identify these errors, provide corrective guidance, and steer the AI toward a successful outcome. This highlights that human expertise acts as the crucial 'final safety net,' ensuring the reliability of AI-generated content. A case involving Deloitte Australia, which submitted a report with fabricated citations to the Australian government, underscores this point; the errors were only detected by a domain expert.
The key to leveraging AI's capabilities lies not in coding skills but in job expertise.
The implications extend beyond coding. The article argues that simply replacing entry-level positions with AI to cut costs is a shortsighted strategy that could stifle long-term organizational competitiveness. Experts like Amy Edmondson of Harvard Business School warn that this approach severs the pipeline for developing future decision-makers. Companies like IBM are rethinking this, increasing entry-level hiring and redesigning roles to integrate AI as a tool, allowing new hires to focus on tasks requiring communication, judgment, and leadership, skills that AI cannot replicate. This strategic integration ensures that human expertise continues to drive innovation and problem-solving.
Communication, judgment, and leadership are essential skills that cannot be replaced by AI.
DistantNews Editorial
Originally published by Dong-A Ilbo in Korean. Translated, summarized, and contextualized by our editorial team with added local perspective. Read our editorial standards.
More Stories
From Our Blog
What Gets Lost (and Found) When News Is Translated
The Critical Role of Diaspora Media in Global News
What Travellers Should Know About Countries with Restricted Press
A Digital Nomad's Guide to Following Local News
How the Same Story Looks Different in Different Countries
