Short answer: both—but timing matters. In the next few years, AI will displace or reshape many tasks faster than it creates entirely new roles. Over longer horizons, if productivity gains lead to new products, lower prices, and business formation, total employment can grow. The net effect depends on how firms deploy AI and how quickly workers and institutions adapt.
What’s most at risk
•Routine, information-heavy tasks: clerical/admin, basic bookkeeping, transcription, call centers, simple content generation, QA/testing, and some entry-level coding and analytics.
•The “junior rung” in professional fields (law, finance, marketing, consulting) where AI can draft, summarize, and analyze—reducing demand for purely executional work.
Where demand is resilient or likely to grow
•Roles needing in‑person presence, dexterity, and reliability: skilled trades, maintenance, manufacturing techs, logistics, hospitality.
•Jobs centered on social intelligence and trust: healthcare and care work, teaching/tutoring, sales, account management, counseling, leadership/management.
•Creative direction, complex problem framing, and cross‑functional project management.
New and expanding roles AI is creating
•AI productization and integration across sectors; MLOps/automation engineers; data governance and privacy; AI safety/evaluation; applied ML in domain contexts.
•“AI augmenters” in every function: marketers, engineers, analysts, designers, and operators who build AI into daily workflows.
What will determine the net outcome
•Complementarity vs substitution: do organizations redesign jobs around human‑AI teams, or mainly cut headcount?
•Diffusion speed and entrepreneurship: do productivity gains translate into new offerings and demand growth?
•Reskilling capacity: access to fast, affordable training and on‑the‑job learning.
•Trust, liability, and regulation: clarity reduces risk and encourages responsible adoption.
•Demographics and sector mix: aging populations and under‑automated services can absorb labor.
How to future‑proof as an individual
•Treat AI as a co‑pilot: learn the leading tools in your field and build repeatable AI‑assisted workflows.
•Double down on human advantages: problem definition, domain expertise, communication, negotiation, leadership, and ethics.
•Build durable tech fluency: data literacy, basic scripting/automation, prompt‑to‑pipeline skills, and an understanding of model limits.
•Show your impact: maintain a portfolio demonstrating how you used AI to save time, improve quality, or open new lines of work.
•Keep learning: short, targeted upskilling every quarter beats one‑off big courses.
What smart employers will do
•Redesign work at the task level; pilot, measure, and scale augmentation before replacement.
•Invest in training and internal mobility; create apprentice‑style pathways into higher‑value tasks.
•Share productivity gains (comp, pricing, investment) to stimulate demand and job growth.
•Put in guardrails: data security, evaluation, and human‑in‑the‑loop for accountability.
Bottom line: AI will both create and take away jobs. Near term, expect more task displacement and job reshaping; medium to long term, net job growth is plausible if we pair adoption with reskilling, job redesign, and policies that turn productivity into new demand. If you tell me your industry and role, I can outline the most likely impacts and a 90‑day upskilling plan.