Compute and infrastructure
Sovereign compute is scarce. Morocco’s TOUBKAL supercomputer and Egypt’s Bibliotheca Alexandrina HPC facility are the exceptions that prove the rule, and fiscal constraints hinder investment elsewhere.
Regional overview · updated from published research
Mediterranean AI: capacity, policy and governance across the region
Artificial intelligence in the Mediterranean is not one story but two. A small group of countries has strategies, data systems and readiness scores that place them within reach of productive AI adoption. A larger group does not. This page sets out what the evidence shows.
Definition
What “Mediterranean AI” means
Mediterranean AI refers to the development, adoption and governance of artificial intelligence across the countries of the Mediterranean basin — the Northern, Southern and Eastern shores taken together rather than as separate blocs.
It is a distinct field of study because the region fits neither a purely Global North nor a purely Global South frame. EU member states on the northern shore are implementing the EU AI Act; countries on the southern and eastern shores are building foundational capacity while adapting global models to Arabic, Amazigh and other regional languages. The governance questions that follow — shared compute, cross-border data, multilingual inclusion, who sets the standards — are regional questions that no national policy answers on its own.
EuroMedAI studies this field as an independent non-profit network. The assessment below is drawn from our own comparative research on the nine South Mediterranean countries where the capacity gap is widest.
The evidence
A two-speed region
Applying the UN Secretary-General’s Tier 0–4 AI maturity framework (A/79/966) to nine South Mediterranean countries, every country assessed sits in the bottom three tiers. None has reached “AI enabled” status. The division within that range is what matters.
- 9
- countries assessed against the UN maturity framework
- 3
- reach Tier 2, “AI ready”: Jordan, Egypt and Morocco
- 0
- have reached Tier 3, “AI enabled”
- 5
- capacity domains measured, from compute to cooperation
Country assessment
AI maturity across the South Mediterranean
Tier classification combines the UN A/79/966 maturity framework with Government AI Readiness Index scores and Open Data Inventory (ODIN) 2024 measures of statistical openness. Palestine is the clearest illustration of why a single score misleads: a strong open-data environment paired with severely limited physical compute and institutional capacity.
| Country | Tier | Classification | Gov AI Readiness | ODIN 2024 |
|---|---|---|---|---|
| Jordan | T2 | AI ready | 61.57 | 72/100 |
| Egypt | T2 | AI ready | 55.63 | 53/100 |
| Morocco | T2 | AI ready | 41.78 | 77/100 |
| Lebanon | T1 | AI experimenter | 46.67 | 33/100 |
| Tunisia | T1 | AI experimenter | 43.68 | 60/100 |
| Algeria | T1 | AI experimenter | 39.06 | 28/100 |
| Palestine | T0 | AI nascent | 37.53 | 75/100 |
| Libya | T0 | AI nascent | 33.25 | 28/100 |
| Syria | T0 | AI nascent | 16.95 | 28/100 |
Source: Bridging the AI Divide (EuroMedAI, 2025) · Read the full study
Diagnosis
What holds the region back
The assessment measures five domains derived from the UN framework. Weakness concentrates in the foundations rather than in ambition: strategy documents exist across the region, but the inputs they depend on frequently do not.
Data and governance
Open-data maturity splits sharply. Morocco (77), Palestine (75) and Jordan (72) score well on ODIN; Algeria, Libya, Lebanon and Syria cluster near 28–33, leaving little machine-readable public data to build on.
Skills and institutions
AI-skilled workforces remain modest and institutional arrangements fragmented. Several countries have begun UNESCO Readiness Assessment Methodology reviews, but ethical AI governance is still being built.
Models and adoption
The region adapts high-capability global models to local linguistic and institutional contexts rather than developing frontier models. Sector pilots in agriculture, health, education and mobility rarely scale past experimentation.
Strategy and cooperation
Uneven connectivity, shallow digital-trade provisions and weak cross-border data governance constrain any scaled regional AI ecosystem, however capable individual countries become.
What would change it
From diagnosis to capacity
Our research proposes measures at two levels. The distinction matters: some constraints are national and budgetary, others cannot be solved by any single country acting alone.
National
- Adopt costed AI strategies rather than unfunded strategy documents
- Invest in data ecosystems and sovereign compute capacity
- Institutionalise ethical AI oversight and human rights safeguards
- Expand AI-relevant skills across the public sector
Regional
- A shared, evidence-based method for measuring AI capacity
- A financed Minimum Irreducible Capacity floor for all states
- A shared South-Mediterranean compute hub
- Cross-border research on Arabic, Darija, Levantine, Tamazight and Kabyle
- A Mediterranean AI Fund to finance the above
Go deeper
The evidence behind this page
Everything summarised here is drawn from published work. The full methodology, country detail and policy argument are in the sources below.
Research and publications
The comparative maturity assessment, the UN Global Dialogue submission, and the methodology behind the tier classifications.
Read the researchProjects
Applied work on advanced-AI safety, inclusive AI literacy in local governance, and AI readiness across the South Mediterranean.
See the projectsEvents
A running record of AI policy and industry convenings across the region — those we host and those we track.
Browse eventsCommon questions
Mediterranean AI, answered
Mediterranean AI refers to the development, adoption and governance of artificial intelligence across the countries of the Mediterranean basin, spanning the Northern, Southern and Eastern shores. It is treated as a distinct field because the region combines EU AI Act implementation on the northern shore with foundational capacity-building on the southern and eastern shores, raising governance questions — shared compute, cross-border data, multilingual inclusion — that no single national policy resolves.
In EuroMedAI’s assessment of nine South Mediterranean countries against the UN Secretary-General’s Tier 0–4 maturity framework (A/79/966), all nine sit in Tiers 0 to 2 and none has reached “AI enabled” status. Jordan, Egypt and Morocco qualify as “AI ready” (Tier 2). Lebanon, Tunisia and Algeria are “AI experimenters” (Tier 1). Palestine, Libya and Syria are “AI nascent” (Tier 0).
Among the nine South Mediterranean countries assessed, Jordan scores highest on the Government AI Readiness Index (61.57), followed by Egypt (55.63) and Lebanon (46.67). On open-data maturity, Morocco leads with an ODIN 2024 score of 77, ahead of Palestine (75) and Jordan (72). Jordan, Egypt and Morocco are the three countries classified as Tier 2, “AI ready”.
The AI divide describes the widening gap in AI capacity between countries, measured across compute, data governance, skills, adoption and strategy. In the South Mediterranean it produces a two-speed trajectory: a small group of “ready but constrained” countries and a larger group of nascent experimenters. EuroMedAI argues that closing this divide is a precondition for legitimate global AI governance, not a secondary concern.
EuroMedAI is an independent non-profit network founded in 2025 in Tirana, Albania, connecting researchers, civil society organisations, policymakers and technologists across Southern Europe, North Africa and the Middle East. It works alongside multilateral processes and institutions including the UN Global Dialogue on AI Governance, UNESCO, ESCWA and the Union for the Mediterranean.
Work on this with us
EuroMedAI is a member network. Researchers, civil society organisations, policymakers and technologists across the region contribute to the evidence base summarised on this page.