01
Professions
Eighteen task bundles already hired: builders, agents, evals, product, risk, literacy, science, climate.
Open the roster →
BITNEWSAI
A BitNews field guide to work in the age of models
BITNE AI
Who does the work when the models show up.
Professions already hired, skills that show up in vacancies, documented deployments, and the reports behind the claims — for BitNews readers who want the labour market, not the launch demo.
Fewer than one percent of workers need advanced model-building skills. Most need to use, question, and supervise the systems.
88%
Organisations using AI in at least one function (McKinsey 2025), up from 78% a year earlier.
~⅓
Vacancies in high AI-exposure occupations — still asking first for management and process skills (OECD 2024).
<1%
Workers who need advanced AI-specific skills such as programming or model development (OECD).
41%
Employees more apprehensive than optimistic — literacy and coaching, not only engineers (McKinsey).
Inside the guide
Six ways in
Start with roles, or skip to skills, workplace literacy, cases, learning paths, or the source list.
02
Skills
Literacy for most workers; model-building for few. Vacancies still ask for management first.
See the clusters →03
Workplace
EU Article 4 literacy measures, SME skill bottlenecks, and what EU vacancies actually show.
Open workplace →04
Examples
Copilot trials, AlphaFold, a heat-stable enzyme study, FDA-cleared imaging, enterprise agents, the EU Skills Academy.
Read the cases →05
Learning
Week-scale literacy, a months-long internal copilot path, workplace measures, and the long specialist route.
Choose a path →06
References
OECD, McKinsey, UNESCO, ESCO, NIST, the EU AI Act, IW Köln, OWASP, and the studies behind the numbers.
Open sources →Preview
Eight of eighteen roles
The rest live on the professions page — including agents, evals, prompt work, safety, conversation design, and AI for science.
Build
AI / ML engineer
Ships models and evaluations. The scarce advanced-skill layer, not the whole labour market.
View roleBuild
MLOps / platform
Turns pilots into production: pipelines, drift, cost, rollback.
View roleProduct
AI product manager
Picks the workflow, the metric, and the human-validation rule.
View roleKnowledge
RAG / knowledge engineer
Citations, access control, stale chunks. Knowledge work is a top surveyed AI function.
View roleRisk
AI ethicist / governance
Inventories, oversight, documentation — process, not posters.
View roleRisk
AI security
Prompt injection, tool agency, model supply chain.
View rolePeople
AI literacy coach
The mass job: most workers need to use and supervise systems.
View roleDomain
Climate / energy AI
Physics plus models. Green work is a labour reallocation; AI is a layer on it.
View roleSources on this page
- McKinsey, The state of AI in 2025.
- OECD skills briefs, 2024–2026.
- McKinsey, Superagency in the workplace, 2025.
- UNESCO AI competency frameworks, 2024.
- European Commission, AI literacy (Article 4) Q&A.
- IW Köln, Fact Sheet #14 on EU AI skill demand, 2026.