Papua New Guinea's Department of Personnel Management piloted an AI system that reads CVs, checks qualifications and drafts shortlists, cutting selection from days to about 30 minutes in a one-month, 200-position trial before a 2025 national rollout.
200+
Positions covered in initial pilot (one-month pilot)
~30 minutes
Shortlisting time (previously days)
Details
Maturity
Pilot
Promoter
Department of Personnel Management (Papua New Guinea)
Period
2025-2026
Keywords
public sector recruitment, human resources, e-government, generative AI
Context
Papua New Guinea's public service has faced criticism for slow, opaque hiring marked by allegations of nepotism. Ahead of the country's 50th independence anniversary in September 2025, the Department of Personnel Management (DPM) began piloting an AI-driven digital recruitment system to replace paper-based files with an end-to-end digital, merit-based process.
Results
The system uses a large language model to read applicants' CVs and cover letters and generate longlists and shortlists automatically. In an initial one-month pilot covering just over 200 positions and a couple of thousand applications, shortlisting exercises that previously took days were completed in about 30 minutes; human interviewers still conduct face-to-face interviews after shortlisting.
Conclusions
No independent audit of the tool's accuracy or bias has been published, and critics have raised concerns about data governance and consent; the government has signalled plans to extend similar AI tools to procurement, tendering and project-delivery oversight.
Implementation
Indicative cost
Low (< €50k)
Time to results
Short (< 1 year)
Staffing & skills
Department of Personnel Management (PNG), Central Agency Coordinating Committee, Departments of Treasury, Finance and National Planning, Justice and Attorney-General, Public Service Commission, Internal Revenue Commission
Conditions for success
human interviewers retained after AI shortlisting
coordination across multiple central government agencies
Common failure modes
no independent audit of accuracy or bias published; data-governance and consent concerns raised by critics
Colombia's ICT Ministry announced six generative-AI models in July 2026 covering citizen complaints, contractor payments, labour certificates and telecom sanctions, …
Baden-Württemberg built F13, the first AI assistant in productive use across a full German state administration, running on sovereign infrastructure; …