Libya’s New AI Strategy Celebrates Participation, Yet Was Built Without It
On June 1, 2026, in a ceremony in Tripoli, the Government of National Unity unveiled what it called a milestone for Libya’s digital future: the National Artificial Intelligence Strategy 2026–2030, alongside a National Charter for AI Ethics. Prime Minister Abdulhamid Dbeibah framed the moment as a turning point: Libya, he said, would build a “knowledge economy” anchored in innovation, digital sovereignty, and “human dignity.” The charter goes further, placing humans at the centre of AI development: systems should support human decision-making, it says, not replace it.
Rights, ethics, and inclusion saturate the vocabulary of both documents. The Ethics Charter lists transparency, fairness, accountability, privacy, and “human values” among its core principles. The strategy promises a “participatory approach,” one that engages government, private sector, and academic stakeholders in shaping the country’s roadmap to 2030. However, in practice, reality has not cooperated.
A read past the main sections of both documents, into their appendices, stakeholder lists, and governance proposals, tells a more complicated story.
Nearly all the consultation behind these texts was conducted inside the state apparatus. Civil society, independent academia, and rights groups were absent from the drafting process. And the institutional design proposed for enforcing AI rights leaves citizens with few independent avenues for redress.
Who Was Actually in the Room?
Thirteen institutions, listed in the strategy’s own appendix, took part in the “national consultations” held in May 2026, filling out the questionnaires underpinning its SWOT analysis. Every one of them is a government ministry, a state-linked regulator, or a state-affiliated company.
Despite “academia” being explicitly named as a partner in the strategy’s own mission statement, there was no university mentioned on the list. No civil society organisation appears, no independent human rights body, no journalists’ or media syndicate, no people with disabilities rights group, no organisation representing women or minorities, and no consumer protection body. The “private sector” voices present are confined to telecom and fintech firms with commercial interests in the policies they helped shape.
None of this is a matter of inference. The strategy states its own process explicitly: the document was prepared, its executive summary says, “through a participatory national approach and consultation with relevant government entities, the private sector, academic institutions, and experts”. Yet the appendix listing actual participants contains none of the latter two categories. The gap between the claim and the documented record is the clearest evidence that “participation,” in this strategy, means inter-ministerial coordination.
Strikingly, the Strategy’s own text seems to know this. Initiative 35, one of 35 in total, calls for “involving the private sector, civil society, and youth in strategy committees.” An annual national forum is proposed under Pillars 1 and 6 as a venue for dialogue with civil society going forward. Read closely, that framing shows that the document treats broad societal engagement as a future deliverable, scheduled for implementation after the strategy has already been adopted, not as a process that informed its content. Civil society is invited to endorse decisions already made, not to help shape them.
The Ethics Charter follows suit. Drafting is attributed to the General Information Authority working “in coordination with sectoral entities and relevant national bodies.” Once again, no civil society actor, no academic institution, and no independent ethics or rights body appear anywhere in the drafting chain.
A Rights Framework Without Independent Teeth
Of the two documents, the Ethics Charter is the more rights-literate. It commits to data privacy, non-discrimination, accountability for algorithmic harm, a ban on the deceptive use of deepfakes, and the principle of meaningful human oversight over high-risk AI decisions in health, justice, and security. These are welcome commitments on paper.
But the institutional architecture meant to enforce them raises serious concerns.
The oversight body is not independent. The charter proposes a “national oversight body for AI ethics” to supervise compliance, approve high-risk systems, and maintain a registry of “ethical compliance” certifications. This body is to be created by, and presumably report within, the Council of Ministers — the same executive structure that owns and operates many of the AI systems it would be auditing. Nothing in the Charter reserves seats for civil society, establishes public nomination processes, requires parliamentary confirmation, or links the body to the judiciary in a way that might insulate it from political interference — a significant omission in a country where the strategy’s own risk register flags “political instability” and “institutional fragmentation” as central threats to implementation, as mentioned in appendix 5.
Redress is described, not engineered. The charter calls for a “clear and accessible mechanism for administrative complaint and objection” to algorithmic decisions, but specifies no appeal timelines, no independent appellate authority, no judicial review pathway, and no obligation to disclose the logic behind an adverse algorithmic decision to the affected person. Context exacerbates this because Libya currently has no functioning data protection authority and no freedom-of-information regime to anchor such a right in practice. Even the strategy concedes as much, noting that the underlying Data Protection and Privacy Law remains in draft form, with adoption targeted only for 2026 under Pillar 2. The Ethics Charter’s privacy commitments, in other words, currently rest on legislation that does not yet exist.
Digital sovereignty is repeatedly paired with individual privacy. Throughout both documents, “protection of personal data” is consistently bundled with “national security” and “digital sovereignty” as a single, undifferentiated value. This framing is not unique to Libya — many national AI strategies invoke sovereignty alongside privacy — but in a context with weak judicial independence and no free press to test such tradeoffs, conflating state security interests with individual rights protections risks subordinating the latter to the former whenever they come into tension.
High-risk classification relies on self-assessment. The charter’s own risk tiering table places justice, security, and health systems in the “high risk” category, requiring “mandatory prior ethical assessment plus human oversight”. Yet the deploying government institution appears to conduct that assessment itself. The oversight body’s role is limited to “periodic audit” after deployment rather than independent prior certification. This is closer to self-regulation than to the independent technical review that comparable frameworks call for, UNESCO’s own Recommendation on the Ethics of AI among them, a document Libya’s own text repeatedly cites as a reference point.
Protections for the most vulnerable remain generic. Persons with disabilities and minors receive a passing mention under the data privacy principle. Neither document addresses risks specific to Libya’s own context: AI’s role in migration and border management affecting refugees and displaced persons, or the risks of biometric surveillance given documented patterns within the security sector. Given that “security” is explicitly listed as a priority deployment sector for AI under both documents, this silence is notable.
Why This Matters
Libya is not alone in producing an AI governance framework that reads better than it was built. Across the region, governments are racing to publish national AI strategies and ethics charters, often explicitly modelled on instruments like UNESCO’s Recommendation or the EU’s regulatory language. The same gap keeps showing up: participatory rhetoric on one hand and closed-door drafting on the other. What distinguishes Libya’s case is how candidly the documents themselves expose this gap: a strategy that names civil society engagement as a future initiative rather than a completed input, and an ethics charter whose enforcement body sits entirely within the executive it is meant to check.
For Libyan civil society organisations, independent technologists, and rights defenders, now is the moment to engage, while the “annual national forum” and the ethics oversight body remain on paper. The strategy itself, perhaps unintentionally, has left the door open by admitting that the participation it claims has not yet actually happened. Whether that door stays open, or whether the 2027 ethics committee and oversight mechanisms are populated the same way the original consultations were — by ministries consulting ministries — will determine whether Libya’s AI governance becomes a genuine rights framework or a procedural shell around an expanding state surveillance and automation apparatus.
This article was published on SMEX on July 22, 2026; Link: https://smex.org/libyas-new-ai-strategy-celebrates-participation-yet-was-built-without-it/.