TUNIS — A restricted ministerial council met on Saturday at the Kasbah, chaired by Prime Minister Sarra Zaafrani Zenzri, to review the state of the country's digital transformation projects. The centrepiece was a presentation by Sofiane Hemissi, Minister of Communication Technologies, of the main lines of a national artificial intelligence strategy covering 2026 to 2030.

The strategy, as presented, rests on three claimed assets: Tunisia's scientific output, its engineering workforce, and its geography. On the first, the council was told the country ranks 29th worldwide in the index of scientific publications in artificial intelligence. On the second, officials pointed to a pool of engineering and academic expertise recognised internationally, and to the Startup Act as a legal framework favourable to innovation. On the third, they described Tunisia's position at the junction of Europe, Africa and the Arab world as the basis for becoming a regional exporter of digital services rather than only a domestic user of them.

Zaafrani Zenzri said the exploitation of the technology had to be accompanied by a preventive approach to its risks, placing data protection and digital sovereignty at the same level as innovation.

The number the council did not mention

Five weeks before the strategy was presented, Tunisia was ranked 110th out of 135 countries in the Global Index on Responsible AI. Business News, reporting that result in July, attributed the position explicitly to the absence of a national strategy — noting that Tunisia trailed both Morocco and Libya.

The two rankings are not contradictory; they measure different things. Twenty-ninth place counts papers. One hundred and tenth place counts institutions: whether a country has enforceable rules on automated decision-making, an independent body able to audit algorithmic systems, procurement standards for AI bought by the state, redress for citizens affected by it.

That gap is the actual diagnosis of Tunisian AI, and it is one the ATUGE white paper published earlier this month reached independently, arguing that the country has talent density without the infrastructure or governance to convert it. A strategy announced in August 2026 has the opportunity to close the second gap. Whether it does depends entirely on what the full text contains — and the full text has not been published.

The rest of the file

The council also reviewed the wider digital programme. According to accounts of the meeting, 192 projects are currently tracked as under way or scheduled for 2026, of which 114 are in progress, several reported at implementation rates above 90 percent. More than 40 public services — including passport applications, driving licences and civil-status documents — are slated to become accessible by mobile.

These are the projects citizens will actually encounter, and they are also where an AI strategy meets its first real test. Digitising a passport application is a service improvement. Automating a decision inside that application — who is flagged, who is delayed — is a governance question, and it is governed by rules that Tunisia, on its own July ranking, does not yet have.

What would close the gap

The measures that international assessments look for are specific and none of them is expensive: a published legal framework for automated decision-making in public administration; an oversight body with the mandate and technical capacity to audit systems the state deploys; procurement rules requiring transparency and impact assessment for AI purchased with public money; and a documented complaints route for citizens.

Tunisia's own National Personal Data Protection Authority already exists and could anchor much of this, as similar authorities have in other jurisdictions. Publishing the strategy's full text, with a timetable and an implementation budget, would be the first verifiable step — and would let the plan be judged on something other than the two rankings it sits between.