On June 21, 2026, The Wall Street Journal published a column arguing that, against the prevailing belief that AI can do almost everything better than humans, AI can do harm in the realms of empathy, authenticity and transparency.
June 21, 2026 · Op-Ed Analysis
What AI Can't — or Shouldn't — Do for You
The belief that AI beats humans at nearly everything is overreach. Forcing it into work that demands empathy, authenticity or transparency can do real harm — a "jagged frontier" where the tool is strong in some places and quietly weak in others.
The historical parallel
Applying AI to every task echoes Project Plowshares — the 1960s plan to repurpose nuclear weapons for peaceful digging that instead spread radioactive fallout. A powerful technology is not a universal one.
15–20
hours a week saved when AI absorbs admin work
300
patients one nurse can manage solo — while staying present in person
Where AI helps vs. where it harms
USE IT — structured load
Transcription & note-taking
Insurance claims
Coding, document review
AVOID IT — human moments
Empathy (a widow canceling an account)
Marketing copy → "AI slop"
Audited decisions: law, payments, medicine
The transparency problem
In regulated fields, today's LLMs act as "biased black boxes."
No way to audit a decision after the fact — explainability remains an open problem.
The over-cut cycle
Pivot heavily to AI
→
Lose accumulated knowledge
→
Rehire & add headcount
Rapid cuts erode institutional knowledge and talent pipelines. The pragmatic middle ground: neither anti-AI nor uncritical cheerleading.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…