42% faster pay growth in one job category. Is it yours?
Your Essential Report from the Frontlines of Applied AI
An important structural change to jobs is underway: the skills that used to only be associated with senior roles, are now expected from juniors.
PwC studied more than 1 billion job ads, and found that in the US, junior roles in jobs most affected by AI are 7 times more likely to require skills such as leadership and creativity.
So which jobs have had their entry bar raised? It is the category of jobs where AI takes over the routine and your judgement matters more; PwC’s examples are radiologists and recruiters. This first category is seeing twice the job growth and 42% faster pay growth.
Comparatively, the second category is more distressing: jobs where AI makes the work easy enough for someone without your expertise to do — IT service managers, medical secretaries. These are the jobs losing ground.
Our Institute’s view is that this shows two unstoppable trends:
- The roles themselves are changing, not just the tasks expected in those roles.
- And the changes will put a higher emphasis on human judgement.
Humans will have to graduate to higher order decisions, and they are more needed than ever.
PwC’s data points the same way: new tasks added to the jobs most exposed to AI are 2.5 times more likely to need empathy, creativity, leadership and judgement.
So ask which category your own job is in, and push it towards the first. If you teach, hire or manage juniors, build those skills in them now — experience will not supply them in time.
The people who build those skills now are the ones who start on the path that pays.
So this week is about what those skills look like in your own work, and how to build them. There’s a prompt below from AI+ TVET Instructors that finds the skills employers now expect but your course does not yet teach, and a case study on IBM, which is tripling its entry-level hiring and rewriting the jobs to match.
Plus the week’s news in applied AI — Denmark’s central bank on who AI-using firms are hiring less, and three places where AI takes over the routine and people keep the decisions: excavator operators in Singapore, GPs in Umbria and case analysts at a Brazilian audit court.
The bar for a first job isn’t disappearing. It’s moving up the skill stack.
So what does a skill higher up the stack actually look like? We split every task into two parts: the how and the what.
The how is the doing: drafting the report, writing the code, building the slides. AI now does much of this, faster than you can.
The what is the deciding: what to make, whether it is worth making, and whether the result is good enough to send. That part stays with you, and it is the part employers now pay for.
So start noticing your own what. Take one piece of work you did with AI this week and write down two things: what the AI did, and what you decided — what to ask for, what you changed, and whether it was good enough.
If you teach or hire juniors, ask them this: ‘Show me a time you disagreed with what the AI suggested, and you were right.’ The ones with an answer are already working higher up the stack.
The more of those decisions you can point to, the higher up the stack you start.
- Labour market
Firms that use AI are hiring slightly less, Denmark’s central bank finds
The slowdown is concentrated among younger and better-educated people. Pay has not moved, and the bank sees no effect on Danish employment overall. Governor Signe Krogstrup said the people who would have been hired found jobs elsewhere.
Danmarks Nationalbank, 16 September 2026
- Construction
Singapore is testing autonomous excavators, with one operator overseeing two machines
JTC Corporation and Kajima launched Singapore’s first pilot of autonomous heavy construction equipment at the Bulim Autonomous Yard, starting with excavators and compactors. Staff monitor the machines remotely and step in when something goes wrong. JTC aims to halve operator headcount.
TNGlobal, 18 September 2026
- Healthcare
GPs in Umbria are trialling an AI that finds the guidelines, while the decision stays with the doctor
About 60 family doctors trained on Mia in Perugia on 16 September. It pulls validated guidelines, treatment pathways and prescribing rules from official sources, cutting the time GPs spend looking them up. The region says responsibility and the decision remain entirely with the doctor. The trial is backed by Agenas, Italy’s national health agency.
Umbria Cronaca, 21 September 2026
- Public audit
A Brazilian audit court gave its case analysts an AI that answers from the court’s own records
Paraíba’s Tribunal de Contas launched Seixas on 16 September, built with the Federal University of Campina Grande. It answers questions from 8.78 terabytes of the court’s archive — 212,362 audit reports and 69,809 decisions among them — and grounds every answer in the documents it drew on.
Resenha Politika, 16 September 2026
Employers raise the bar faster than qualification standards move. This prompt compares the two: what employers increasingly expect that is not in your standard yet, and what is in it that is losing relevance. It is written for vocational training and works for any course or programme you teach — swap the trade for your subject. Keep its last line: check every suggestion against your own industry knowledge and your employers.
I teach [your trade] at [level]. The current qualification standard covers these competency outcomes: [list outcomes]. Based on current industry trends in [your trade/industry], identify: (1) skills or competencies that employers increasingly expect but are not yet in the qualification standard, (2) technologies or methods that are becoming industry-standard but may not be reflected in training materials, and (3) competency outcomes in the current standard that may be declining in relevance. For each gap, suggest how I could address it within the existing programme structure — through supplementary content, enrichment activities, or employer-led workshop sessions. Note: I will verify all suggestions against my own industry knowledge and employer advisory board feedback.
Prompt 4.3 from AI+ TVET Instructors, word for word.
Keep hiring juniors, and rebuild their jobs around the work that needs a person.
In February, IBM said it would triple its US entry-level hiring in 2026. Its chief human resources officer, Nickle LaMoreaux, was clear about which jobs: “yes, that is for software developers and all these jobs we’re being told AI can do.”
The jobs themselves are new. LaMoreaux told a New York summit that AI can do most of the entry-level jobs of two or three years ago, so IBM rewrote them. Its new software developers spend less time on routine coding and more time with customers, working out what they actually want built. Its new HR staff refine the answers from AI chatbots, which now handle more enquiries than the staff could answer themselves.
Her reason is the pipeline. Cut junior hiring now, she argued, and in a few years you run short of mid-level managers — and the experienced people you hire from competitors instead cost more and take longer to fit in.
If you hire juniors, write the job around the part that needs judgement from day one. If you teach them, that is the part they will be hired for.