For each country, we aim to identify:
- The ministry/department responsible for AI policy
- Named contacts (ministers, secretaries, advisors)
- Contact information (official emails, not personal)
- Summit participation history
- Whether they have a national AI Safety Institute
Search patterns:
"[Country] Ministry of Technology AI policy""[Country] Ministry of Digital Affairs""[Country] Ministry of Economy digital strategy""[Country] national AI strategy office"
Common ministry names by region:
- Europe: Ministry of Digital Affairs, Ministry of Economy, Ministry of Science
- Asia: Ministry of Science and ICT (Korea), Ministry of Economy (Japan), MeitY (India)
- Americas: Varies widely - often split across multiple agencies
- MENA: Often Ministry of Communications or dedicated digital authority
Priority order:
- Minister/Secretary (political appointee)
- Permanent Secretary/Deputy (senior civil servant)
- Director of AI/Digital Policy
- Head of AI Safety Institute (if exists)
- GPAI national delegate (if findable)
Search patterns:
"[Country] [Ministry name] AI policy director""[Country] AI Safety Summit 2024 delegation""[Country] GPAI delegate representative"
Check declaration signatory lists:
- Bletchley 2023: gov.uk declaration page
- Seoul 2024: Search for "Seoul AI Safety Summit declaration signatories"
- Paris 2025: Search for "Paris AI Action Summit statement signatories"
Record:
signed- Signed the declarationattended- Attended but didn't signdeclined- Explicitly declinedabsent- Did not participateunknown- Can't determine
Search: "[Country] AI Safety Institute" OR "[Country] national AI safety"
Known institutes (as of Jan 2026):
- UK: AI Safety Institute (AISI)
- US: US AI Safety Institute (USAISI) - under NIST, status uncertain post-Trump
- Japan: AI Safety Institute Japan
- Canada: Canadian AI Safety Institute (announced 2024)
- France: Part of INRIA
- Singapore: AI Safety workstream under IMDA
Priority sources (in order):
- Official government website contact pages
- Ministry press office contacts
- Published consultation response addresses
- OECD/GPAI listed contacts
Do not use:
- Personal email addresses
- LinkedIn scraped data
- Unofficial sources
Every piece of information needs:
- Source URL
- Date accessed
- Confidence level (high/medium/low)
Use data/countries/_template.md - copy and fill in for each country.
Before marking a country as "complete":
- Lead ministry identified with source
- At least one named contact with title
- Official contact email (even if generic ministry address)
- Summit participation for all three summits checked
- AI Safety Institute status confirmed
- All sources documented with dates
- Information less than 12 months old
Record what you searched, mark as "needs_escalation" in PROGRESS.md. Some countries genuinely don't have dedicated AI policy structures.
List primary ministry first, note others in the country file. Common for larger countries to split AI across tech, economy, and security ministries.
Note the date of last election/transition. Political appointees change; permanent secretaries usually don't.
Include if they participated in summits (e.g., China, UAE at Bletchley). Use same methodology.
Some data can be partially automated:
- OECD AI dashboard has structured policy data
- GPAI member list is stable and published
- Summit signatory lists are in structured documents
Scripts in /scripts handle these where possible. Manual research fills gaps.