A pre-apprenticeship should earn progression, not attendance
A UK review proposes a national pre-apprenticeship route for young people outside work and education. The first contract should define the next step and measure who reaches it.
Topic desk / 04
Skills Systems and HR Tech
How are skills data, inference, assessment, mobility, learning, and workforce-planning systems changing?
← All latest reporting60 articles
A shared employability language still needs observable evidence
Skills England’s new framework embeds AI use across 13 employability skills. Providers should translate each statement into a task, trace and assessor rule before treating it as evidence.
Prompting is becoming a baseline; hiring needs evidence of applied AI work
Workday reports that demand for basic AI skills fell after a January peak while demand for building and automation skills rose. The useful response is a work-sample ladder, not a new keyword list.
Teen AI planning needs verifiable hand-offs, not another persistent helper
OpenAI announced college-planning and study features for ChatGPT for Teens as an independent assessment reported safeguard failures. Education buyers need to test when the system hands control to a student, parent or professional.
A cheaper small model needs task-specific error budgets before high-volume rollout
Anthropic positions Claude Haiku 5.5 for classification, extraction, routing and support at materially lower cost. Scale changes the risk equation: buyers need error budgets by task, not one benchmark average.
SAP’s TechWolf deal makes skills-data lineage a deployment decision
SAP plans to acquire TechWolf and connect its work, task and skills graph to SuccessFactors. Before inferred skills shape mobility or learning, buyers need lineage, freshness and contestability rules.
Anthropic’s 10,000-engineer target needs project evidence, not a badge count
Claude Frontier Academy combines a graded simulation with a 12-week workplace project. Employers should define the project record now, before the 10,000-person target becomes the outcome.
AI verification does not preserve junior lawyers’ practice by itself
A federal judge paired citation checking with a warning about stunted training. Legal teams need protected unaided drafting and supervised revision, not only a final accuracy checklist.
Positive tutoring effects do not make an AI tutor implementation-ready
A European Commission systematic review finds the strongest evidence for learning outcomes, with more mixed motivation evidence and little socio-emotional evidence. Procurement should therefore test the local learning mechanism, not import an average effect.