Oracle laid off 30,000 people. Goldman Sachs warned that "AI will automate 25% of US working hours." Go by headlines alone and it sounds like AI is set to swallow every job on the planet. But at the same time, a WSJ report analyzing LinkedIn data tells a completely different story — from 2023 to 2025, AI created 640,000 new jobs in the US alone. And that's before counting a single data center construction job.
What Is It?
On April 2, 2026, WSJ reporter Te-Ping Chen published an analysis of LinkedIn hiring data. The headline: "Wanted: Head of Human AI Solutions. The New Jobs Being Created by AI." It's a data-driven counterargument to a conversation that's been almost entirely focused on jobs AI destroys.
Here's what the report found:
LinkedIn's chief economist Kory Kantenga put it this way: "It's not a scale that changes the entire direction of the labor market — but AI-related job growth is literally going vertical."
And this isn't just a US story. The WEF's Future of Jobs Report 2025 shows the bigger picture — by 2030, 170 million new jobs will be created globally while 92 million disappear. Net gain: 78 million. AI isn't simply taking jobs away — it's redrawing the entire playing field.
What Changes?
So specifically — which roles are disappearing, and which ones are being created?
| Roles Shrinking | New Roles Emerging |
|---|---|
| Basic coding & QA testing | Full-stack AI engineer / GenAI engineer |
| Structured data entry & processing | Data annotator (312,000 new roles) |
| Basic translation & transcription | AI trainer (by specialty, $70–$200/hr) |
| General administrative support | Head of AI / AI strategy lead (225,000 roles, +49%) |
| Rules-based customer service chatbot management | AI governance & ethics specialist |
| Basic legal document review | Head of Human AI Solutions (human-AI collaboration design) |
| Basic medical imaging reads | AI-assisted medical imaging specialist (hiring up 35%) |
See the pattern? What's shrinking: "tasks that are repetitive and rule-bound." What's growing: "roles that require judgment, context, and specialized knowledge."
Josh Bersin's analysis nails it. Looking at software engineering hiring data, he says — "AI doesn't eliminate jobs — it eliminates tasks. And when those tasks disappear, the job itself gets richer." Software engineer salaries have actually risen more than 15% in the AI era.
The real people profiled in WSJ's piece make this vivid:
Daniel Millian (42, pathologist) — By day, he reads biopsy slides at a hospital. After hours, he puts in 4–5 hours as an AI trainer: writing virtual medical scenarios and evaluating AI model responses. Rate: $90–$200/hr. Side income last year alone: $75,000.
Zach Kinzler (25, BoodleBox) — Fresh out of grad school, he landed the title "Head of Human AI Solutions." His core skill isn't coding — it's "the ability to explain technology to non-experts." He openly admits he's learning on the fly.
Victoria Chapa (32, formerly Meta) — She does short-term gig work: describing the emotional impact of AI-generated images and training chatbots to better match user tone. Her honest take: "It's not a good job. It pays the bills, but staring at AI images all day makes you feel like you're losing your mind." She's now pivoting toward AI governance and ethics.
The spectrum across these three cases matters. The jobs AI creates range from high-earning side work for PhD-level specialists to a new kind of management role to unstable gig work. "AI is creating jobs" doesn't automatically mean "everything is getting better."
Getting Started: How to Position Your Career in the AI Era
Recurring reports, data cleanup, basic translation, scheduling — be honest about what percentage of your workday these take up. According to Goldman Sachs, 46% of tasks in administrative support, 44% in legal, and 37% in architecture and engineering are automatable by AI.
As micro1 CEO Ali Ansari puts it, "As AI takes on increasingly complex specialized tasks, it will continuously need human experts to train it." That's why pathologists, financial professionals, and lawyers are the most expensive 'teachers' for AI models.
WSJ's profile of Zach Kinzler found his core skill wasn't coding — it was "the ability to explain AI to non-experts." AI literacy is becoming a career in its own right. As of late 2025, only 6% of companies are posting AI-related roles — we're still in the early innings.
This is what Josh Bersin calls the "Superworker." Hiring of medical imaging specialists jumped 35% not in spite of AI — but because AI handling initial diagnostics expanded the need for patient care, data management, and communication. As technical automation rises, so does the market value of human skills.
Some AI-created jobs pay $200/hr for expert training work. Others mean sorting AI-generated images all day. Oracle laying off 30,000 people and AI startups paying new grads $300K+ are happening at the same time. Knowing where you actually stand — honestly — is step one.
Deep Dive Resources
- WSJ Original Article — Te-Ping Chen's full piece. LinkedIn data analysis with real-person interviews.
- WEF Future of Jobs Report 2025 — Survey of 1,000+ global companies across 55 countries and 22 industries. Job change projections through 2030.
- Josh Bersin's Analysis — Detailed job-by-job data on software engineers, medical imaging specialists, and more — plus the "Superworker" framework.
- Goldman Sachs Research — AI automation rates by profession (admin support 46%, legal 44%, architecture & engineering 37%).
- Washington Post — A breakdown of 16 new job roles AI has created (Knowledge Architect, Orchestration Engineer, and more).
