Africa AI Governance Index 2026 · Country profile on AI governance

Country Profile on AI Governance: São Tomé and Príncipe

São Tomé and Príncipe scores 0.92 out of 4 on the Africa AI Governance Index 2026. It sits in the Nascent tier and ranks 36 of 54 states.

Region
Central Africa
Version
1.2 · 7 October 2026
Evidence reference date
30 June 2026
Overview

São Tomé and Príncipe at a glance

How to read this profile

The Index is diagnostic. It measures formal governance instruments, not capability. Composite differences under about 0.1 are not meaningful.

Country

Population (2024)
235,536
Region
Central Africa
Regional economic communities
Economic Community of Central African States
Geography
Coastal, 2 international submarine cable systems in service

Governance status

  • National AI strategy S1.1

    1 / 4

  • Data protection law G2.2

    3 / 4

  • AI regulatory body G2.5

    1 / 4

  • Malabo Convention ratification G2.7

    3 / 4

  • Kigali Declaration endorsement R7.2

    3 / 4

Scores are indicator scores on the 0–4 key in the method section, not maturity tiers.

Headline assessment

São Tomé and Príncipe scores 0.92 out of 4 on the Africa AI Governance Index 2026. It sits in the Nascent tier and ranks 36 of 54 states. The score rests on 79 of 80 indicators scored from the fellow country profile. The strongest pillar is Governance & Regulation (1.50). The weakest is Innovation & Ecosystem (0.20). The Central Africa mean is 1.07. The mean for all 54 states is 1.33.

Strongest pillar
Governance & Regulation1.50 / 4Emerging
Weakest pillar
Innovation & Ecosystem0.20 / 4Nascent

Indicators scored: 79 / 80

Pillar scores

Pillar scores at a glance

Each pillar is scored from 0 to 4 as the mean of its scored indicators.

Bar chart of São Tomé and Príncipe's eight pillar scores out of 4, compared with the Central Africa mean and the mean of all 54 states. The same figures are listed beside each bar and in the table below.
  • São Tomé and Príncipe score
  • Central Africa mean
  • All 54 states mean
  1. P1Strategy & Vision
    1.20Emerging

    Central Africa: 1.40
    All 54 states: 1.76

  2. P2Governance & Regulation
    1.50Emerging

    Central Africa: 1.31
    All 54 states: 1.32

  3. P3Infrastructure & Data
    1.00Emerging

    Central Africa: 1.20
    All 54 states: 1.51

  4. P4Human Capital
    0.30Nascent

    Central Africa: 0.69
    All 54 states: 1.20

  5. P5Innovation & Ecosystem
    0.20Nascent

    Central Africa: 0.67
    All 54 states: 1.05

  6. P6Ethics & Inclusion
    1.30Emerging

    Central Africa: 1.16
    All 54 states: 1.26

  7. P7Regional Integration
    1.20Emerging

    Central Africa: 1.34
    All 54 states: 1.55

  8. P8Implementation & Impact
    0.50Nascent

    Central Africa: 0.60
    All 54 states: 0.78

  9. Composite (weighted)
    0.92Nascent

    Central Africa: 1.07
    All 54 states: 1.33

View the pillar scores as a table
São Tomé and Príncipe pillar scores on the Africa AI Governance Index 2026
PillarBasisWeightScore (0–4)MaturityCentral Africa meanAfrica mean
P1 Strategy & VisionMean of 10/10 indicators15%1.20Emerging1.401.76
P2 Governance & RegulationMean of 10/10 indicators15%1.50Emerging1.311.32
P3 Infrastructure & DataMean of 9/10 indicators15%1.00Emerging1.201.51
P4 Human CapitalMean of 10/10 indicators15%0.30Nascent0.691.20
P5 Innovation & EcosystemMean of 10/10 indicators10%0.20Nascent0.671.05
P6 Ethics & InclusionMean of 10/10 indicators10%1.30Emerging1.161.26
P7 Regional IntegrationMean of 10/10 indicators10%1.20Emerging1.341.55
P8 Implementation & ImpactMean of 10/10 indicators10%0.50Nascent0.600.78
Composite (weighted)—100%0.92Nascent1.071.33
Country report

What the scores say about São Tomé and Príncipe

São Tomé and Príncipe scores 0.92 out of 4 on the Africa AI Governance Index 2026, in the Nascent tier, and ranks 36 of 54. Fellows scored 79 of 80 indicators. The strongest pillar is Governance & Regulation at 1.50 and the weakest is Innovation & Ecosystem at 0.20. The composite is below the Central Africa mean of 1.07 and the Africa mean of 1.33. The evidence explains the profile through a 2016 data protection law with an operating agency, ratification of the Malabo Convention in 2023 and a World Bank digital project. The country has no AI strategy, no AI law, no confirmed commercial data centre and no AI companies, researchers or training programmes on record.

P1Strategy & Vision1.20 / 4Emerging

Strategy & Vision scores 1.20, in the Emerging tier, with 10 of 10 indicators scored. No standalone national AI strategy is adopted (S1.1, score 1; S1.2, score 0). The evidence points to adjacent instruments: a National Cybersecurity Strategy 2024 to 2028 launched in December 2023, a National Strategy for Digital Governance and a National Development Plan 2025 to 2039. Political commitment scores 2 on ministerial engagement with digital transformation, with no AI-specific commitment formalised (S1.3). The country has no ring-fenced AI budget; the World Bank Digital STP Project had disbursed USD 7.71 million of USD 21.4 million by March 2025 (S1.4, score 1). Alignment with the African Union Continental AI Strategy is not confirmed (S1.5, score 1).

See the 10 indicators for Strategy & Vision

P2Governance & Regulation1.50 / 4Emerging

Governance & Regulation scores 1.50, in the Emerging tier, with 10 of 10 indicators scored. Law No. 03/2016 on the Protection of Personal Data is enacted, and the National Data Protection Agency has operated since 2018 (G2.2, score 3). The country ratified the Malabo Convention on 25 September 2023 and deposited its instrument on 15 February 2024 (G2.7, score 3). The agency has no explicit AI oversight mandate, and its staff count and budget are not disclosed (G2.3, G2.4, score 2 each). No dedicated AI regulatory body exists (G2.5, score 1). No AI-specific legislation is enacted or under formal discussion (G2.1, score 0), and intellectual property law has not been reviewed for AI (G2.10, score 0).

See the 10 indicators for Governance & Regulation

P3Infrastructure & Data1.00 / 4Emerging

Infrastructure & Data scores 1.00, in the Emerging tier, with 9 of 10 indicators scored. The index counts 2 submarine cable systems (I3.3, score 2). Internet penetration stood at 61.5% in January 2025, with 146,000 users (I3.4, score 2). Law No. 03/2016 requires prior authorisation from the National Data Protection Agency for cross-border transfers (I3.7, I3.8, score 2 each). No Tier-rated commercial data centre is confirmed, and no capacity figure is published (I3.1, I3.2, score 0 each). No hyperscaler has local infrastructure (I3.5, score 0), and no high-performance computing facility is confirmed (I3.6, score 0). Foreign AI surveillance spending is a context flag and is not scored; none is documented (I3.9).

See the 10 indicators for Infrastructure & Data

P4Human Capital0.30 / 4Nascent

Human Capital scores 0.30, in the Nascent tier, with 10 of 10 indicators scored. Seven indicators score 0. No AI researchers affiliated with national institutions were identified (H4.1), and no AI or machine learning degree programme is confirmed at the University of São Tomé and Príncipe (H4.3). No AI patents are filed (H4.5). No labour transition framework or informal economy assessment exists (H4.8, H4.9). The Creole languages Forro, Angolar and Principense have zero coverage in language technology (H4.10). Three indicators score 1. The National Cybersecurity Strategy calls for a cybersecurity curriculum in schools and universities and for training of IT professionals, with no mandatory computing curriculum and no AI reskilling programme confirmed (H4.6, H4.7). Publications score 1 (H4.4).

See the 10 indicators for Human Capital

P5Innovation & Ecosystem0.20 / 4Nascent

Innovation & Ecosystem scores 0.20, in the Nascent tier, with 10 of 10 indicators scored. Eight indicators score 0. The evidence confirms no AI companies (N5.1), no private AI investment (N5.2), no AI incubators or accelerators (N5.3) and no regulatory sandbox (N5.4). The country has no AI safety testing facility (N5.6), no ISO/IEC 42001 adoption (N5.7), no standards body with AI technical capacity (N5.8) and no university and industry AI partnerships (N5.10). Two indicators score 1. Government digital procurement is ad hoc, with no AI procurement framework (N5.5). The data protection agency's notification and authorisation process works as a partial risk assessment for high-risk data processing, with no AI-specific framework (N5.9).

See the 10 indicators for Innovation & Ecosystem

P6Ethics & Inclusion1.30 / 4Emerging

Ethics & Inclusion scores 1.30, in the Emerging tier, with 10 of 10 indicators scored. Law No. 03/2016 gives data subjects rights to information, access, rectification and erasure (E6.3, score 2). The National Data Protection Agency has run a complaint mechanism since 2018 (E6.4, score 2). The agency belongs to the Network of African Data Protection Authorities (E6.7, score 2). A European Union Civil Society Roadmap 2023 to 2027 documents civil society engagement, with no permanent AI co-governance (E6.8, score 2). No formal algorithmic impact assessment requirement exists (E6.1, score 1). No formal AI advisory body or youth mechanism exists (E6.6, E6.10, score 1 each). The evidence finds no bias audit requirements (E6.2, score 0).

See the 10 indicators for Ethics & Inclusion

P7Regional Integration1.20 / 4Emerging

Regional Integration scores 1.20, in the Emerging tier, with 10 of 10 indicators scored. Endorsement of the Africa Declaration on AI scores 3 on the basis of the signatory list (R7.2). African Union alignment scores 2 through Malabo Convention ratification, with no confirmed engagement in African Union AI work (R7.1). Regional coordination scores 2 through ECCAS and CPLP membership, with no AI-specific coordination confirmed (R7.3). No delegation to the July 2026 Geneva Dialogue is confirmed (R7.9, score 1). Three indicators score 0: no provisions address AI in elections (R7.5), no framework covers AI-generated disinformation (R7.6), and government has no formal engagement with platforms on content moderation (R7.8).

See the 10 indicators for Regional Integration

P8Implementation & Impact0.50 / 4Nascent

Implementation & Impact scores 0.50, in the Nascent tier, with 10 of 10 indicators scored. Budget disbursement scores 2: the World Bank Digital STP Project had disbursed USD 7.71 million, or 36%, of its IDA grant by March 2025 (M8.1). Three indicators score 1. Financial services have no AI governance framework and no central bank AI guidance (M8.7). The education evidence covers cybersecurity curriculum plans, with no AI education policy (M8.8). Recorded outcomes concern digital governance, not AI (M8.9). Six indicators score 0: health AI regulation (M8.2), health AI device approval (M8.3), a government AI system registry (M8.4), government AI monitoring reports (M8.5), agriculture AI governance (M8.6) and adoption of World Health Organization guidance on AI for health (M8.10).

See the 10 indicators for Implementation & Impact

Indicator evidence

Evidence by pillar

Evidence and score for each of the 80 indicators, with sources. Scores use the 0–4 indicator score key in the method section, which is a different scale from the maturity tiers. “Not scored” means nothing could be determined, or a score had no supporting evidence: the indicator is left out of the pillar mean rather than counted as 0. “N/A” and “Context only” sit outside the score.

P1 · Weight 15% of composite

Strategy & Vision

1.20 / 4Emerging

Mean of all 10 indicator scores. Central Africa mean 1.40 · all 54 states mean 1.76.

  1. S1.1

    National AI strategy existence

    Score 1 / 4

    No standalone national AI strategy adopted; National Development Plan 2025-2039 references technology and digital economy as cross-cutting themes; National Strategy for Digital Governance (developed with UNU support) addresses digital public administration; National Cybersecurity Strategy 2024-2028 (launched December 2023) includes AI-adjacent digital governance; World Bank Digital STP Project (P177158, effective February 2023) supports modernization of telecom legal and regulatory framework; no AI-specific policy adopted

    Note: No AI-specific strategy; digital governance strategy and cybersecurity strategy serve as indirect AI instruments; scored 1 as nascent, digital strategy framework exists without AI policy

    Sources dig.watch (opens in a new tab)

  2. S1.2

    Strategy adoption date

    Score 0 / 4

    No AI strategy adoption date; National Cybersecurity Strategy 2024-2028 was launched December 2023; National Development Plan 2025-2039 adopted 2024 but does not constitute an AI strategy

    Note: Not applicable until an AI-specific strategy is formally adopted

    Scoring basis Threshold applied. Raw value: none (adoption date). No formally adopted national AI strategy.

  3. S1.3

    Political commitment signals

    Score 2 / 4

    Prime Minister and Minister of Planning engaged in National Development Plan 2025-2039 development; World Bank partnership (Country Partnership Framework FY2024-2029, endorsed January 2024) signals executive-level commitment to digital transformation as one of five CPF pillars; National Cybersecurity Strategy launched at ministerial level (December 2023); blockchain-based citizenship programme (CBI) launched October 2025 with UAE partner IOPn signals executive-level digital innovation appetite

    Note: Ministerial-level engagement on digital transformation documented; no presidential AI directive; AI-specific commitment not yet formalised; scored 2 as formal institutional actions with multi-year commitment

    Sources worldbank.org (opens in a new tab)

  4. S1.4

    Dedicated AI budget allocation

    Score 1 / 4

    No ring-fenced AI budget; World Bank Digital STP Project (IDA grant) has disbursed USD 7.71 million (36%) by March 2025 of total USD 21.4 million; National Development Plan references digital economy but no AI-specific budget line; national budget heavily donor-dependent (ODA approximately 70% of public investment)

    Note: Digital investment through World Bank IDA grant; AI-specific budget absent; scored 1 as nascent, embedded digital budget without AI allocation

    Sources documents1.worldbank.org (opens in a new tab)

  5. S1.5

    AU Continental AI Strategy alignment

    Score 1 / 4

    AU member and ECCAS member; National Development Plan 2025-2039 references international development frameworks; no confirmed formal alignment with AU Continental AI Strategy; Africa Declaration on AI (Kigali, April 2025) was signed by 49 countries but STP's individual endorsement not confirmed from public sources; CPLP (Portuguese Language Community) digital cooperation framework applies

    Note: Passive AU alignment through membership; no active AU AI alignment process; scored 1 as awareness-level

  6. S1.6

    Multi-stakeholder consultation

    Score 2 / 4

    World Bank Digital STP Project developed through multi-stakeholder process including government, private sector, and civil society; National Cybersecurity Strategy 2024-2028 involved stakeholder consultation; UNU National Strategy for Digital Governance included public consultation; EU Civil Society Roadmap 2023-2027 documents civil society engagement in digital governance

    Note: Formal consultation documented through World Bank and digital strategy processes; permanent ongoing mechanism not institutionalised; scored 2 as formal consultation with institutional record

    Sources dig.watch (opens in a new tab)

  7. S1.7

    Sector prioritisation

    Score 2 / 4

    National Cybersecurity Strategy 2024-2028 identifies priority areas for digital security; National Development Plan 2025-2039 identifies tourism, agriculture, fisheries, and digital economy as priority sectors; World Bank Digital STP Project focuses on telecom modernisation and data governance; no AI-specific sector prioritisation

    Note: Digital sector priorities established; AI-specific sector roadmaps absent; scored 2 as digital priorities identified without AI focus

    Sources worldbank.org (opens in a new tab)

  8. S1.8

    Implementation roadmap

    Score 1 / 4

    National Cybersecurity Strategy 2024-2028 has a phased implementation roadmap (four pillars, 2024-2028); World Bank Digital STP Project has milestones through June 2027; National Development Plan 2025-2039 provides a 14-year development roadmap; no AI-specific implementation plan

    Note: Digital and cybersecurity roadmaps exist; AI-specific implementation plan absent; scored 1 as nascent, sector roadmap without AI milestones

    Sources dig.watch (opens in a new tab)

  9. S1.9

    Monitoring and evaluation framework

    Score 1 / 4

    World Bank Digital STP Project includes M&E framework with Implementation Status Reports (latest: September 2025 mission); National Cybersecurity Strategy includes monitoring provisions; no AI-specific M&E framework

    Note: World Bank project M&E active; sector-level digital monitoring; AI-specific M&E absent; scored 1

    Sources documents1.worldbank.org (opens in a new tab)

  10. S1.10

    Strategy review mechanism

    Score 1 / 4

    World Bank Digital STP Project has built-in review mechanism through Implementation Status Reports and closing date of June 2027; National Cybersecurity Strategy 2024-2028 has annual review provisions; no formal AI strategy review mechanism

    Note: External review through World Bank project cycle; no adaptive AI governance mechanism; scored 1

P2 · Weight 15% of composite

Governance & Regulation

1.50 / 4Emerging

Mean of all 10 indicator scores. Central Africa mean 1.31 · all 54 states mean 1.32.

  1. G2.1

    AI-specific legislation

    Score 0 / 4

    No AI-specific legislation enacted; National Cybersecurity Strategy 2024-2028 calls for updating laws on cybercrime, data protection, and digital identity; World Bank Digital STP Project supports modernisation of telecom legal framework; no AI law under formal discussion

    Note: Pre-legislative stage on AI; cybersecurity strategy calls for legal updates; scored 0

  2. G2.2

    Data protection law status

    Score 3 / 4

    Law No. 03/2016 on the Protection of Personal Data enacted; ANPDP (National Data Protection Agency) established and operational since 2018; Law No. 7/2017 on Organisation and Operation of ANPDP; cross-border transfers require prior ANPDP authorisation; administrative fines of USD 1,250-50,000; independently structured DPA with state budget funding; World Bank Digital STP Project supporting creation of integrated data protection management platform

    Note: Law enacted 2016; ANPDP operational since 2018; enforcement fines active; independently structured; World Bank platform support ongoing; scored 3 as established, operational DPA with enforcement

    Sources dataprotection.africa (opens in a new tab)

  3. G2.3

    DPA AI mandate

    Score 2 / 4

    ANPDP operational since 2018 with mandate covering all data processing including automated decisions; cross-border transfer provisions require ANPDP authorisation; ANPDP is member of NADPA-RAPDP (Network of African Data Protection Authorities); Smart Africa-NADPA MoU signed to advance data protection enforcement and harmonisation; no explicit AI oversight mandate confirmed

    Note: ANPDP well-established DPA; AI-relevant mandate through automated processing provisions; no explicit AI mandate; NADPA participation indicates capacity building; scored 2 as developing, DPA with relevant mandate

    Sources rapdp.org (opens in a new tab)

  4. G2.4

    DPA staffing and budget

    Score 2 / 4

    ANPDP independently structured (confirmed by dataprotection.africa); receives direct state budget funding; has adopted multiple resolutions (July 2018) covering specific processing exemptions; specific staff count and annual budget not publicly disclosed; ANPDP participates in NADPA-RAPDP capacity-building workshops

    Note: ANPDP operational with confirmed independent structure and state funding; staff/budget not disclosed; NADPA capacity building ongoing; scored 2 as developing, operational without disclosed capacity metrics

    Scoring basis Fellow score kept; no raw value for the threshold.

    Sources dataprotection.africa (opens in a new tab)

  5. G2.5

    AI regulatory body existence

    Score 1 / 4

    No dedicated AI regulatory body; ANPDP is the data protection authority; National Cybersecurity Strategy establishes a Cybersecurity Committee and CERT-STP (National Cybersecurity Incident Response Centre), both in development; no AI-specific governance body

    Note: General ICT/cybersecurity governance developing; no AI-specific body; Cybersecurity Committee under formation; scored 1 as nascent, AI-adjacent bodies being established

    Sources dig.watch (opens in a new tab)

  6. G2.6

    Institutional design model

    Score 2 / 4

    Emerging multi-body approach: ANPDP (data protection), Cybersecurity Committee (under formation per 2024-2028 Strategy), CERT-STP (under development), CSTII (state ICT authority); World Bank Digital STP Project modernising regulatory framework; data governance policy expected October 2025

    Note: Multiple bodies developing; data governance policy under creation; scored 2 as developing, institutional design actively being constructed

    Sources documents1.worldbank.org (opens in a new tab)

  7. G2.7

    Malabo Convention ratification

    Score 3 / 4

    São Tomé and Príncipe ratified the AU Malabo Convention on 25 September 2023; instrument deposited 15 February 2024, confirmed on Wikipedia Malabo Convention page (row 47: STP signed 29/01/2016, ratified 25/09/2023, deposited 15/02/2024); STP is among the 16 ratifiers; Malabo ratification is notable for a small island developing state

    Note: ADVANCED for context, Malabo ratified September 2023, deposited February 2024; full ratification obligations active; notably among the most recently ratified; scored 3 as established, ratified with deposits confirmed

    Sources en.wikipedia.org (opens in a new tab)

  8. G2.8

    Enforcement actions taken

    Score 1 / 4

    ANPDP has been operational since 2018 but no specific AI-related enforcement actions documented; general data protection enforcement through ANPDP notification and authorisation process; National Cybersecurity Strategy 2024-2028 targets enhanced judicial capacity for cybercrime; no AI enforcement precedent

    Note: ANPDP enforcement capacity exists through fine regime (USD 1,250-50,000); no AI-specific enforcement; scored 1 as nascent, enforcement capacity established without AI precedent

    Sources dataguidance.com (opens in a new tab)

  9. G2.9

    Extraterritorial provisions

    Score 1 / 4

    Law No. 03/2016 applies to processing within STP national territory; cross-border transfer restrictions require ANPDP authorisation; no confirmed extraterritorial scope for foreign processors of STP data; narrower than GDPR model

    Note: Territorial-only scope with outbound transfer controls; no confirmed extraterritorial provisions for inbound processing; scored 1 as nascent

    Sources dataguidance.com (opens in a new tab)

  10. G2.10

    IP framework for AI

    Score 0 / 4

    No review of IP legislation for AI-generated content identified; ANPDP does not have an AI IP mandate; no IP law amendment process confirmed

    Note: IP framework not reviewed for AI; scored 0

P3 · Weight 15% of composite

Infrastructure & Data

1.00 / 4Emerging

Mean of the 9 scored indicators; 1 carries no score. Central Africa mean 1.20 · all 54 states mean 1.51.

No score: 1 context only.

  1. I3.1

    Data center count

    Score 0 / 4

    No Tier-rated commercial data centre confirmed in São Tomé and Príncipe; government ICT infrastructure (CSTII) hosts basic servers; World Bank Digital STP Project supporting data governance infrastructure; digital connectivity remains nascent with most government systems on externally-hosted cloud services; FactSheet records 'internet speeds 5-50 Mbps urban' suggesting local hosting capacity exists but not at commercial Tier level

    Note: Basic government ICT infrastructure exists; no Tier-rated commercial data centre; World Bank project improving governance infrastructure; scored 1 as nascent

    Scoring basis Threshold applied. Raw value: 0 (data centres). Raw value read from the evidence: "No Tier-rated commercial data centre confirmed in São Tomé and Príncipe" [absence of confirmed commercial DC]. Checked against the evidence on 7 October 2026: confirmed.

    Sources saotomeexpert.pt (opens in a new tab)

  2. I3.2

    Data center capacity (MW)

    Score 0 / 4

    No published MW capacity data; government data centre capacity not disclosed

    Note: Infrastructure pre-commercial Tier; cannot quantify; scored 0

    Scoring basis Fellow score kept; no raw value for the threshold.

  3. I3.3

    Submarine cable landing points

    Score 2 / 4

    ACE (Africa Coast to Europe) submarine cable connected since 2012, transformative for STP connectivity; World Bank Digital STP Project is procuring a submarine cable connection to Príncipe island (separate from São Tomé island) to address inter-island connectivity gap; Príncipe cable procurement facing delays as of September 2025 mission; 1 confirmed international submarine cable landing (ACE on São Tomé)

    Note: ACE cable confirmed operational since 2012; Príncipe cable under procurement (delayed); single cable dependency risk; scored 2 as developing, international connection established with expansion underway

    Scoring basis Threshold applied. 2 cable systems (TeleGeography, accessed 2026-10-06).

    Sources ts2.tech (opens in a new tab)

  4. I3.4

    International bandwidth (Gbps)

    Score 2 / 4

    Internet penetration 61.5% as of January 2025 (146,000 users, DataReportal); 71.4% mobile connection rate (170,000 connections); speeds 5-50 Mbps urban; above Africa average (~39%); however single ACE cable creates vulnerability; high broadband costs; Príncipe island remains less connected; rural-urban digital divide persists

    Note: 61.5% penetration is well above African average; ACE cable enabled transformation; single cable dependency and high costs constrain quality; scored 2 as developing, meaningful penetration with structural constraints

    Scoring basis Fellow score kept; no raw value for the threshold.

    Sources datareportal.com (opens in a new tab)

  5. I3.5

    Hyperscaler presence

    Score 0 / 4

    No confirmed AWS, Azure, Google Cloud, or other hyperscaler local infrastructure; island geography and small market make hyperscaler local investment unlikely; STP uses externally hosted cloud services for government systems

    Note: No hyperscaler has committed to STP; scored 0

  6. I3.6

    GPU/HPC access

    Score 0 / 4

    No HPC or GPU facility confirmed; University of São Tomé and Príncipe (USTP) has very limited computing resources; no AI research computing infrastructure

    Note: Small island economy constrains compute infrastructure; scored 0

  7. I3.7

    Data localisation provisions

    Score 2 / 4

    Law No. 03/2016 includes data transfer restrictions: cross-border transfers require ANPDP prior authorisation unless adequate protection confirmed; National Data Revolution equivalent policy referenced in World Bank project; 2017 National Data Revolution Policy requires critical government data locally hosted (World Bank ISR references data governance policy under development)

    Note: Law 03/2016 cross-border transfer controls operational; ANPDP authorisation required; developing data governance policy; scored 2 as developing, legal framework operational

    Sources dataguidance.com (opens in a new tab)

  8. I3.8

    Cross-border data flow framework

    Score 2 / 4

    Law No. 03/2016 restricts cross-border transfers to countries with adequate protection or with ANPDP authorisation; Malabo Convention ratification provides regional framework; ANPDP active in NADPA cross-border cooperation; data governance policy expected October 2025 will strengthen framework

    Note: Legal cross-border framework operational; Malabo provides regional multilateral framework; governance policy strengthening in progress; scored 2 as developing

    Sources dataprotection.africa (opens in a new tab)

  9. I3.9

    Foreign AI surveillance spending

    Context only: Context flag outside the score

    No documented government spending on foreign AI surveillance systems; STP's small scale and democratic governance context make AI surveillance procurement unlikely; no evidence found

    Note: No evidence found; scored 0

    Scoring basis Context flag; outside the score. Raw value read from the evidence: "No documented government spending on foreign AI surveillance systems".

  10. I3.10

    Sovereign cloud initiatives

    Score 1 / 4

    World Bank Digital STP Project explicitly targets 'creation of an integrated data protection management platform' and data governance infrastructure; data governance policy expected October 2025; no confirmed operational sovereign cloud but active investment in government data infrastructure

    Note: Sovereign data infrastructure investment active through World Bank project; policy under development; not yet operational sovereign cloud; scored 1 as nascent, plan and investment confirmed

    Sources worldbank.org (opens in a new tab)

P4 · Weight 15% of composite

Human Capital

0.30 / 4Nascent

Mean of all 10 indicator scores. Central Africa mean 0.69 · all 54 states mean 1.20.

  1. H4.1

    AI researchers per million

    Score 0 / 4

    No AI researchers affiliated with STP institutions identified; University of São Tomé and Príncipe (USTP) has very limited STEM capacity; population of approximately 241,000 makes per-million count effectively zero; Digital Watch Observatory AI Readiness data infrastructure score: 40.39/100

    Note: No identifiable AI researchers; very small population and limited university capacity; scored 0

    Scoring basis Threshold applied. Raw value: 0 (AI researchers per million). Raw value read from the evidence: "makes per-million count effectively zero" [no AI researchers identified]. Checked against the evidence on 7 October 2026: confirmed.

    Sources dig.watch (opens in a new tab)

  2. H4.2

    DPA staff with AI competency

    Score 0 / 4

    ANPDP operational but AI-specific staff competency not published; small DPA with limited disclosed staffing

    Note: ANPDP operational since 2018; AI competency data not available; scored 0

    Scoring basis Fellow score kept; no raw value for the threshold.

  3. H4.3

    University AI programs

    Score 0 / 4

    University of São Tomé and Príncipe (USTP) offers general degree programmes; National Cybersecurity Strategy 2024-2028 calls for development of national cybersecurity curriculum in schools and universities; no confirmed AI or ML degree programme; UNICEF Learning Passport digital platform integrated in schools; World Bank education project supports skills development

    Note: University infrastructure exists; cybersecurity curriculum under development; no AI programme; scored 1 as nascent, education institutions exist without AI curriculum

    Scoring basis Threshold applied. Raw value: 0 (universities). Raw value read from the evidence: "no confirmed AI or ML degree programme". Checked against the evidence on 7 October 2026: confirmed.

    Sources dig.watch (opens in a new tab)

  4. H4.4

    African-authored AI publications

    Score 1 / 4

    No AI-related academic publications from STP-affiliated authors identified; very small academic sector

    Note: Academic sector too small to generate AI output; scored 0

    Scoring basis Threshold applied. 1.0 papers/yr ÷ 0.24m people = 4.25 per million (floors applied). Raw value read from the evidence: "No AI-related academic publications from STP-affiliated authors identified" [source used: not named]. Checked against the evidence on 7 October 2026: confirmed.

  5. H4.5

    AI patents filed

    Score 0 / 4

    No AI-related patents filed by STP-affiliated inventors

    Note: Patent activity absent; scored 0

    Scoring basis Threshold applied. Raw value: 0 (AI patent applications/year).

  6. H4.6

    K-12 CS education

    Score 1 / 4

    National Cybersecurity Strategy 2024-2028 explicitly calls for developing a national cybersecurity curriculum in schools and universities and increasing public awareness and digital literacy; UNICEF Learning Passport digital platform in schools; World Bank Social Protection and Skills Development Project supports education; no mandatory AI/CS curriculum confirmed

    Note: Digital literacy referenced in cybersecurity strategy and education projects; no mandatory CS/AI curriculum; scored 1 as nascent, digital literacy priority identified

    Sources dig.watch (opens in a new tab)

  7. H4.7

    AI workforce training programs

    Score 1 / 4

    National Cybersecurity Strategy 2024-2028 calls for training IT professionals and general users on cyber hygiene; establishing national and international partnerships for cybersecurity training; ANPDP participates in NADPA capacity-building workshops; no government-funded AI-specific reskilling programme confirmed

    Note: Cybersecurity training planned; ANPDP capacity building active; no AI workforce programme; scored 1 as nascent

    Sources dig.watch (opens in a new tab)

  8. H4.8

    Labour transition policy

    Score 0 / 4

    No formal government framework for AI and labour market transition; labour market is small (~33,000 total workforce per World Bank); no assessment commissioned

    Note: Very small labour market; AI transition not a current policy concern; scored 0

  9. H4.9

    Informal economy AI assessment

    Score 0 / 4

    No government assessment of AI's impact on the informal economy; informal sector significant but no AI-specific assessment

    Note: Scored 0

  10. H4.10

    African language NLP coverage

    Score 0 / 4

    Portuguese is the official language, one of the best-covered languages in global LLMs (Portuguese is a top-10 language in most large language models including GPT-4 and Gemini); Forro, Angolar, and Principense Creoles spoken domestically have zero NLP coverage; no domestic AI language tool development confirmed

    Note: Portuguese official language coverage through global models is a structural advantage; Creole languages (Forro, Angolar, Principense) have zero NLP coverage; scored 1 as Portuguese NLP coverage benefits STP through global tools

    Scoring basis Threshold applied. Raw value: 0 of 3 (languages supported / major languages), normalised 0. 0% of major domestic languages.

P5 · Weight 10% of composite

Innovation & Ecosystem

0.20 / 4Nascent

Mean of all 10 indicator scores. Central Africa mean 0.67 · all 54 states mean 1.05.

  1. N5.1

    AI companies count

    Score 0 / 4

    Very nascent tech ecosystem; most digitally advanced entity is the blockchain-based Citizenship by Investment (CBI) programme (launched October 2025 with UAE partner IOPn); GTI Bank (Ghanaian consortium, launched February 2025) entering with digital services focus; no AI companies confirmed; e-Saotome digital payment platform reported as growing

    Note: Nascent but emerging digital sector; blockchain CBI and digital banking are most advanced deployments; no dedicated AI companies; scored 1

    Scoring basis Threshold applied. Raw value: 0 (companies). Raw value read from the evidence: "no AI companies confirmed". Checked against the evidence on 7 October 2026: confirmed.

    Sources coingeek.com (opens in a new tab)

  2. N5.2

    Private AI investment (USD)

    Score 0 / 4

    No private AI investment documented; FDI remains limited; GTI Bank (digital services, February 2025) is the most significant recent digital investment but not AI-specific

    Note: Investment climate constrained by small market size; no AI investment; scored 0

    Scoring basis Threshold applied. Raw value: 0 (USD/year). Raw value read from the evidence: "No private AI investment documented". Checked against the evidence on 7 October 2026: confirmed.

  3. N5.3

    AI incubators/accelerators

    Score 0 / 4

    No AI incubators or technology accelerators confirmed; no general tech hub documented in São Tomé city

    Note: Innovation support infrastructure does not exist; scored 0

    Scoring basis Threshold applied. Raw value: 0 (incubators/accelerators). Raw value read from the evidence: "No AI incubators or technology accelerators confirmed". Checked against the evidence on 7 October 2026: confirmed.

  4. N5.4

    Regulatory sandbox (AI)

    Score 0 / 4

    No AI regulatory sandbox identified; no fintech sandbox despite digital banking growth

    Note: Scored 0

  5. N5.5

    Government AI procurement

    Score 1 / 4

    Government actively engaged in blockchain CBI programme (October 2025); World Bank Digital STP Project is a significant government digital procurement; national e-government modernisation under way; no formal domestic AI procurement preference policy

    Note: Ad hoc government digital procurement active; no AI-specific procurement framework; scored 1

    Sources coingeek.com (opens in a new tab)

  6. N5.6

    AI safety evaluation capacity

    Score 0 / 4

    No AI safety testing or evaluation facility; ANPDP has no AI safety mandate; CERT-STP being established under cybersecurity strategy but not yet operational

    Note: Scored 0

  7. N5.7

    ISO/IEC 42001 adoption

    Score 0 / 4

    No ISO/IEC 42001 adoption; national standards body capacity extremely limited

    Note: Scored 0

  8. N5.8

    Standards body AI competency

    Score 0 / 4

    No confirmed national standards body with AI technical capacity

    Note: Scored 0

  9. N5.9

    AI risk assessment frameworks

    Score 1 / 4

    ANPDP notification and authorisation process constitutes a de facto risk assessment for high-risk data processing; National Cybersecurity Strategy 2024-2028 includes identifying and regulating critical infrastructure; no AI-specific risk assessment framework

    Note: ANPDP notification as partial proxy risk assessment; cybersecurity critical infrastructure identification planned; no AI risk framework; scored 1

    Sources dig.watch (opens in a new tab)

  10. N5.10

    University-industry AI partnerships

    Score 0 / 4

    No formal university-industry AI partnerships; USTP has no documented private sector AI collaboration

    Note: Scored 0

P6 · Weight 10% of composite

Ethics & Inclusion

1.30 / 4Emerging

Mean of all 10 indicator scores. Central Africa mean 1.16 · all 54 states mean 1.26.

  1. E6.1

    Algorithmic impact assessment

    Score 1 / 4

    ANPDP notification process requires data controllers to notify of processing activities including automated decisions; no formal Algorithmic Impact Assessment requirement; data protection impact assessments not explicitly required under Law 03/2016 but ANPDP authorisation for high-risk processing serves a similar function

    Note: ANPDP notification serves as partial AIA proxy; no standalone AIA requirement; scored 1 as nascent

    Sources dataguidance.com (opens in a new tab)

  2. E6.2

    Bias audit requirements

    Score 0 / 4

    No bias audit requirements; no regulatory guidance on AI bias

    Note: Scored 0

  3. E6.3

    Transparency obligations

    Score 2 / 4

    Law No. 03/2016 establishes data subject rights including right to information about processing, right of access, rectification, and erasure; ANPDP can enforce transparency obligations; National Cybersecurity Strategy references transparency and trust as guiding principles

    Note: Data law transparency provisions operational through ANPDP; cybersecurity strategy references transparency principle; no AI-specific transparency obligation; scored 2 as developing, operational legal transparency framework

    Sources dataprotection.africa (opens in a new tab)

  4. E6.4

    Citizen redress mechanism

    Score 2 / 4

    ANPDP complaint mechanism operational since 2018; data subjects can file complaints with ANPDP about processing violations including automated decisions; administrative fines of USD 1,250-50,000 applicable; judicial recourse through national courts available

    Note: ANPDP complaint mechanism operational with confirmed fine regime; judicial recourse available; no dedicated AI complaint channel; scored 2 as developing, operational DPA complaint mechanism

    Sources dataprotection.africa (opens in a new tab)

  5. E6.5

    Judicial AI remedy

    Score 1 / 4

    National courts adjudicate disputes; ANPDP provides administrative review before judicial challenge; no documented AI-specific court cases; Cybercrime Law (Law 17/2017) provides judicial framework for digital offences

    Note: Courts and ANPDP administrative review available; no AI-specific precedent; scored 1 as nascent

    Sources lawgratis.com (opens in a new tab)

  6. E6.6

    AI advisory body diversity

    Score 1 / 4

    No formal AI advisory body; ANPDP has a Board of Directors with some multi-sector representation; National Cybersecurity Strategy proposes a Cybersecurity Committee (not yet operational); EU Civil Society Roadmap 2023-2027 documents civil society space for digital governance

    Note: General data protection governance body (ANPDP Board) exists; AI advisory body absent; scored 1 as nascent

  7. E6.7

    Global AI governance participation

    Score 2 / 4

    AU member; ECCAS member; ANPDP is member of NADPA-RAPDP (Network of African Data Protection Authorities), active participation in African DPA network; Smart Africa-NADPA MoU participation; Malabo Convention ratified; invited to accede to Budapest Convention (October 2023); Digital Watch Observatory tracks STP in AI governance; no confirmed GPAI engagement

    Note: Active NADPA participation; Malabo ratified; Budapest Convention invitation; no GPAI membership; scored 2 as active observer in African AI governance networks

    Sources rapdp.org (opens in a new tab)

  8. E6.8

    Civil society AI engagement

    Score 2 / 4

    EU Civil Society Roadmap for STP 2023-2027 documents formal civil society engagement mechanisms; FONG-STP (Federation of NGOs) active in governance monitoring; World Bank Digital STP Project included civil society consultation; ANPDP consultation processes documented

    Note: Formal civil society engagement through EU roadmap and World Bank process; FONG-STP active; no permanent AI co-governance; scored 2 as developing

    Sources eeas.europa.eu (opens in a new tab)

  9. E6.9

    African ethical framework reference

    Score 1 / 4

    National Cybersecurity Strategy 2024-2028 references human rights and freedoms as a guiding principle including privacy and freedom of expression; no explicit reference to African ethical frameworks (Ubuntu, etc.); Portuguese-language cultural context

    Note: Human rights reference in cybersecurity strategy; no African philosophical grounding; scored 1 as mentioned in cybersecurity context

    Sources dig.watch (opens in a new tab)

  10. E6.10

    Youth participation

    Score 1 / 4

    National Cybersecurity Strategy 2024-2028 explicitly targets youth through school curricula development; UNICEF Learning Passport in schools; no formal youth AI governance mechanism; STP has a young population (median age approximately 19 years)

    Note: Youth digital education referenced; no formal AI governance youth channel; scored 1

    Sources dig.watch (opens in a new tab)

P7 · Weight 10% of composite

Regional Integration

1.20 / 4Emerging

Mean of all 10 indicator scores. Central Africa mean 1.34 · all 54 states mean 1.55.

  1. R7.1

    AU AI Strategy alignment

    Score 2 / 4

    AU member; Malabo Convention ratified (September 2023, deposited February 2024), aligns with AU cybersecurity and data protection framework; National Development Plan 2025-2039 references international development frameworks; no confirmed active engagement in AU AI Technical Committee

    Note: Malabo ratification represents formal AU governance alignment; no active AU AI strategy engagement; scored 2 as developing, alignment process via Malabo ratification

    Sources en.wikipedia.org (opens in a new tab)

  2. R7.2

    Kigali Declaration endorsement

    Score 3 / 4

    STP participation in Africa Declaration on AI (Kigali, April 2025) not individually confirmed; declaration signed by 49 African countries but STP-specific endorsement not separately documented

    Note: Likely signatory as AU member but not individually confirmed; scored 1

    Scoring basis Official signatory list applied. Signatory of the Africa Declaration on AI (Kigali, 4 April 2025), C4IR list: "Endorsed" = 3.

    Sources c4ir.rw (opens in a new tab)

  3. R7.3

    REC AI governance coordination

    Score 2 / 4

    ECCAS member; CPLP (Community of Portuguese Language Countries) digital cooperation framework; Central African digital cooperation programmes referenced; ANASP (national telecom regulator) engages with regional bodies; no confirmed AI-specific ECCAS coordination

    Note: ECCAS and CPLP membership provide regional digital frameworks; AI-specific coordination not confirmed; scored 2 as active regional participant

    Sources africanexponent.com (opens in a new tab)

  4. R7.4

    Cross-border AI agreements

    Score 2 / 4

    Malabo Convention ratification (September 2023, deposited February 2024) provides multilateral data governance framework; National Cybersecurity Strategy 2024-2028 calls for international cooperation for cybercrime enforcement; invited to accede to Budapest Convention (October 2023); ANPDP NADPA-RAPDP cross-border cooperation

    Note: Malabo ratification provides signed multilateral framework; Budapest Convention invitation pending accession; NADPA cross-border cooperation active; scored 2

    Sources en.wikipedia.org (opens in a new tab)

  5. R7.5

    Election integrity AI provisions

    Score 0 / 4

    No specific legal or institutional provisions addressing AI in elections identified; National Electoral Commission exists; STP has a democratic multi-party system since independence; no AI elections governance

    Note: Democratic system stable; no AI elections framework; scored 0

  6. R7.6

    Disinformation response framework

    Score 0 / 4

    No formal government framework for AI-generated disinformation; National Cybersecurity Strategy 2024-2028 addresses cybercrime and digital trust but not specifically AI disinformation

    Note: Cybersecurity strategy covers digital trust broadly; no AI disinformation framework; scored 0

  7. R7.7

    AI election incidents documented

    Score 1 / 4

    No publicly documented AI-related election interference incidents

    Note: Stable democracy; no AI election incidents; scored 0

    Scoring basis Threshold applied. Raw value: 1 (rule band). Scored on documented institutional response, not incident count. Raw value read from the evidence: "National Cybersecurity Strategy 2024-2028 addresses cybercrime and digital trust" [incidents: 0]. Checked against the evidence on 7 October 2026: confirmed.

  8. R7.8

    Platform content moderation

    Score 0 / 4

    No formal government engagement with social media platforms on AI content moderation for STP context; very small social media population (~68,800 user identities, January 2025)

    Note: Very small social media population; platform governance not a priority; scored 0

  9. R7.9

    Geneva Dialogue participation

    Score 1 / 4

    No confirmed STP delegation to UN Global Dialogue on AI Governance (July 2026); small state diplomatic capacity limits multilateral AI governance participation; STP active in UN system through CDP and UNCTAD processes

    Note: Limited diplomatic capacity for specialised AI multilateral engagement; scored 1 as observer-level through UN system

  10. R7.10

    G20/multilateral AI engagement

    Score 1 / 4

    Not a G20 or BRICS member; active in AU, ECCAS, CPLP, and UN multilateral digital frameworks; Malabo Convention and NADPA provide AI-adjacent governance engagement; no active AI governance leadership role

    Note: Observer-level in NADPA and Malabo multilateral AI governance; no leadership role; scored 1

    Sources rapdp.org (opens in a new tab)

P8 · Weight 10% of composite

Implementation & Impact

0.50 / 4Nascent

Mean of all 10 indicator scores. Central Africa mean 0.60 · all 54 states mean 0.78.

  1. M8.1

    AI governance budget disbursed

    Score 2 / 4

    World Bank Digital STP Project (P177158) has disbursed USD 7.71 million (36%) of IDA grant by March 2025; USD 0.28 million (35%) of GIF grant disbursed; total USD ~22 million IDA/GIF package active; data governance policy and strategy expected October 2025; CERT-STP under development; telecom legal modernisation in progress

    Note: USD 7.71M confirmed disbursed; multiple digital governance outputs under development; scored 2 as developing, significant disbursement with governance outputs in progress

    Scoring basis Fellow score kept; no raw value for the threshold.

    Sources documents1.worldbank.org (opens in a new tab)

  2. M8.2

    Health AI regulation

    Score 0 / 4

    No health AI regulation; no AI-specific provisions in health sector; Ministry of Health operates with WHO support but no AI health governance

    Note: Pre-regulatory stage; scored 0

  3. M8.3

    Health AI devices approved

    Score 0 / 4

    No AI health devices or tools formally approved; no regulatory mechanism for health AI

    Note: Scored 0

    Scoring basis Threshold applied. Raw value: 0 (authorised devices).

  4. M8.4

    Government AI system registry

    Score 0 / 4

    No government registry of AI systems in use; e-government systems nascent

    Note: Scored 0

  5. M8.5

    AI governance M&E reports

    Score 0 / 4

    World Bank Digital STP Project Implementation Status Report (September 2025 mission) documents project progress; no government AI governance M&E report; ANPDP has published resolutions (July 2018) as governance documentation

    Note: World Bank ISR as external governance assessment; no government AI M&E report; scored 1

    Scoring basis Threshold applied. Raw value: 0 (reports).

    Sources documents1.worldbank.org (opens in a new tab)

  6. M8.6

    Agriculture AI governance

    Score 0 / 4

    No AI governance provisions for agriculture; cocoa, copra, and fishing are primary agricultural activities; no AI agricultural deployment documented

    Note: Scored 0

  7. M8.7

    Financial services AI governance

    Score 1 / 4

    BCSTP (Central Bank of São Tomé and Príncipe) governs financial services; GTI Bank (launched February 2025) explicitly targeting digital services and financial inclusion but no AI governance framework; no central bank AI guidance; digital payments growing but no AI-specific financial governance

    Note: Digital banking growing with GTI Bank launch; no AI financial governance; scored 1 as nascent, digital financial growth creates governance pressure

    Sources state.gov (opens in a new tab)

  8. M8.8

    Education AI governance

    Score 1 / 4

    National Cybersecurity Strategy 2024-2028 explicitly includes school and university cybersecurity curriculum development; UNICEF Learning Passport digital platform integrated in schools; World Bank Skills Development Project active; no formal AI education policy adopted

    Note: Cybersecurity education in strategy; digital platforms in schools; no AI education policy; scored 1

    Sources dig.watch (opens in a new tab)

  9. M8.9

    Measurable governance outcomes

    Score 1 / 4

    Blockchain-based CBI programme (October 2025) is a documented government digital governance innovation; World Bank Digital STP Project has documented governance outcomes (USD 7.71M disbursed, CERT-STP development, data governance policy); ANPDP has been operational for 8 years with resolution adoption; ACE submarine cable (2012) enabled internet penetration growth to 61.5%, a measurable digital governance outcome

    Note: Anecdotal digital governance outcomes: CBI blockchain, World Bank project outputs, ANPDP resolutions, ACE cable impact; no AI governance outcomes; scored 1

    Sources coingeek.com (opens in a new tab)

  10. M8.10

    WHO AI health guidance adoption

    Score 0 / 4

    No formal adoption of WHO Ethics and Governance of AI for Health guidance; WHO operates health programmes in STP but no AI health governance alignment

    Note: Scored 0

Method

What is the Africa AI Governance Index?

The Africa AI Governance Index 2026 assesses 8 pillars of national AI governance. Each pillar is examined through 10 indicators. The Index is diagnostic. It measures formal governance instruments, not capability. Composite differences under about 0.1 are not meaningful.

Indicator score key

Indicator score key, 0 to 4
ScoreDefinition
0 No evidenceNo policy, institution, or activity. Confirmed absence.
1 NascentInitial discussion, concept note, informal working group.
2 DevelopingDraft policy, institution forming, pilot, or formal consultation.
3 EstablishedAdopted policy or operational institution; limited enforcement.
4 AdvancedFully operational with enforcement and measurable outcomes.

Maturity tiers

Applied to pillar and composite means.

Maturity tiers by score range
TierMean score
NascentBelow 1.00
Emerging1.00 to 1.99
Developing2.00 to 2.99
Established3.00 and above

How scores are built

A pillar score is the mean of the scored indicators in that pillar. Confirmed absence scores 0. An indicator where nothing could be determined is excluded from the pillar mean and reported as not scored. A score with no supporting evidence is excluded.

The composite is 15% of each of pillars P1 to P4 plus 10% of each of pillars P5 to P8, on a scale of 0 to 4. In other words, P1, P2, P3, P4 carry 15% each and P5, P6, P7, P8 carry 10% each.

Source: Africa AI Governance Index 2026 Final Dataset, 54 countries, built 6 October 2026. Lawyers Hub, Africa AI Policy Lab. Version 1.2, 7 October 2026.

Suggested citation

Lawyers Hub, Africa AI Policy Lab (2026). Country Profile on AI Governance: São Tomé and Príncipe. Africa AI Governance Index 2026, version 1.2, 7 October 2026. https://www.aipolicy.africa/country-profiles/sao-tome-and-principe