The OECD.AI Policy Navigator

Our policy navigator is a living repository from more than 80 jurisdictions and organisations. Use the filters to browse initiatives and find what you are looking for.

Use of AI - National Observatory of Strategic Foresight


Added by:   OECD analyst
Added on:   16 Nov 2026
Updated by:   OECD analyst
Updated on:   16 Nov 2026

Peru’s National Strategic Planning Center (Ceplan) uses AI (GPT-4o) to enhance its National Foresight Observatory. The initiative, implemented by the National Directorate of Strategic Foresight and Studies (DNPE), uses AI as a methodological assistant for editing, data analysis and standardising technical files. Human oversight is mandatory to ensure accuracy. The goal is to improve efficiency and consistency in identifying future scenarios without replacing expert decision-making.

Name in original language

Uso de la IA - Observatorio Nacional de Prospectiva Estratégica

Initiative overview

The National Directorate of Strategic Foresight and Studies (DNPE) of Ceplan—the governing body of Peru’s National Strategic Planning System (Sinaplan) attached to the Presidency of the Council of Ministers—is leading an initiative to incorporate Artificial Intelligence (AI) into its technical analysis. This project uses AI chatbots (specifically GPT-4o) as methodological and editorial assistants to enhance the efficiency and quality of the National Observatory of Strategic Foresight. The AI supports the specialised team in processing technical sheets that cover conceptualisation, drivers, and strategic recommendations.

Key Applications: 

  • a. Standardisation: Ensuring uniform structure across technical sheets. 
  • b. Editorial Improvement: Enhancing grammar, technical style, and conceptual precision. 
  • c. Data Extraction: Recovering text from images and charts. 
  • d. Monitoring: Automating news alerts and identifying relevant context signals. 
  • e. Content Updates: Suggesting recent evidence and references for validation. 
  • f. Analytical Support: Assisting with Stata, Python, or R code for data processing.

Mandatory Human Oversight: A core principle of this initiative is that AI acts as a support tool, not an autonomous decision-maker. All AI-generated outputs are strictly reviewed and validated by the technical team to ensure veracity and methodological soundness.

Technical Specifications: 

  • a. Model: GPT-4o (OpenAI).
  • b. Capabilities: Multimodal LLM with advanced contextual reasoning and structured JSON output. 
  • c. Selection Criteria: Chosen for its stability in strategic classification and superior performance in complex comparative analysis.

Results, outcomes and impacts

Ceplan’s National Directorate of Strategic Foresight has achieved concrete results by integrating AI into its technical production: a. Standardisation: Improved consistency and comparability of "future events" technical sheets. b. Efficiency: Reduced time spent on editorial review and quality control, shifting focus to expert validation. c. Data Recovery: Enhanced conversion of images and scanned PDFs into editable text. d. Monitoring: Optimised daily tracking of emerging signals and news through automated, uniform reporting.

Other relevant details

Challenges and lessons learned: The integration of AI at CEPLAN has highlighted significant challenges, such as the risk of "hallucinations" (plausible but false data), a heavy reliance on prompt quality, and a tendency towards "wordiness" with low technical value. Additionally, AI often struggles with broken links and unverified sources, requiring rigorous manual fact-checking. Key Lessons Learned: Human Oversight is Non-negotiable: AI is a support tool, not a substitute. Specialist critical judgement remains the core element to ensure veracity and analytical depth. Garbage In, Garbage Out: The utility of AI is directly tied to the precision of instructions. Success requires institutional methodologies and explicit standards (fixed structures and technical tones). Focus on Process, Not Decisions: The greatest gains are found in automating editorial tasks and systematisation rather than delegating strategic analysis. Validation over Autonomy: AI-generated outputs are contrasted with reliable sources.

About the policy initiative


Category:

  • AI policy initiatives, programmes and projects

Initiative type:

  • AI use cases/projects in the public sector

Status:

  • Active

Start Year:

  • 2025

OECD AI Principles:

—

Other relevant urls:

—