Why regional
No country in the region can evaluate frontier models, negotiate access conditions or respond to cross-border incidents on its own.
The Forum
why now
Frontier models and agentic AI mark a leap in innovation, but they also open up cybersecurity and sustainability risks. Systems no longer just respond: they plan, use tools and carry out tasks with growing autonomy.
Context
Innovation
Productivity, public services, scientific research and more competitive SMEs.
Cybersecurity
The same capability that speeds up defence also speeds up attack. It makes it easier to discover vulnerabilities, carry out social engineering and automate intrusions. When an agent causes harm across jurisdictions, it is still unclear how liability is allocated and how the response is coordinated.
International AI Safety Report, 2025
Sustainability
Electricity demand from data centres will more than double by 2030.
International Energy Agency, 2025
Central America faces this double edge with small response teams, recent regulatory frameworks and high dependence on external providers. That is why the response has to be regional.
The region is not starting from scratch
Under the Regional Digital Strategy (ERDI), SICA is moving forward with AI and cybersecurity projects and with a Charter of ethical principles for AI. Sectoral authorities already hold applicable powers, so the priority is to speed up their coordination, not to create new bodies.
What the Forum seeks
A high-level dialogue for Central America, together with Latin America and the Caribbean, to set shared priorities on frontier and agentic AI towards 2027.
Diagnose
Risks and opportunities in cybersecurity, critical infrastructure, energy, water and development.
Align
Principles for adopting agents in government and a common position towards large technology providers.
Project
A regional agenda that feeds SICA's AI Strategy and reaches the UN Second Global Dialogue.
Outcome
forum conclusions
They follow the four pillars of the Global Dialogue:
- Pillar 1
AI opportunities for development
Productivity, public services, SMEs and strategic sectors.
- Pillar 2
Closing gaps
Compute, data, connectivity, talent and equitable access to international cooperation.
- Pillar 3
Safe and trustworthy AI
Frontier models, autonomous agents, cross-border incidents and critical infrastructure.
- Pillar 4
Human rights, transparency and accountability
Human oversight, traceability and interoperability between national frameworks.
Four principles
The diagnostic findings come together around one idea: use what already exists before creating something new.
Use the institutions that already exist
Sectoral regulators, current rules and incident reporting before creating new bodies.
Regulate effects, not technologies
Verifiable harms and liability for those who develop or operate agents; insurance as an incentive.
Act at the speed AI demands
Control and safety are an immediate priority; sovereignty and infrastructure, a medium-term priority.
Negotiate as a region
Common positions on procurement and jurisdiction, and partnerships with other regions through international organisations.
Co-organisers
who makes it possible
MICITT
Host: content, technical roundtables, drafting of the conclusions and launches
Ministry of Foreign Affairs and Worship of Costa Rica
Diplomatic outreach and liaison with the Permanent Mission of Costa Rica to the United Nations
SICA
Regional outreach; links the conclusions to the Regional Digital Strategy (ERDI) and the Regional AI Strategy
United Nations System in Costa Rica
Link to the Global Dialogue process and technical input from its agencies
British Embassy in Costa Rica
AI Futures Outlook keynote and Tuesday keynote
Clúster Cybersec
Cybersecurity industry; moderates the cybersecurity panel
CAMTIC
ICT industry; moderates the Industry Vision panel