Africa AI Ledger · News
West African Election Authorities Validate AI Guidelines
Election officials from across West Africa have reviewed regional guidelines intended to govern AI use while protecting electoral integrity, transparency and public trust.

Background generated in ChatGPT; Africa AI Ledger editorial template
West African election authorities are working toward a shared set of guidelines for the responsible use of artificial intelligence in election administration, moving a regional debate about emerging technology closer to practical standards for electoral bodies.
Representatives met in Abuja on 12 and 13 October to validate draft guidelines developed under the Network of Electoral Commissions of ECOWAS Member States, known as ECONEC. According to the National Elections Commission of Liberia, the discussions focused on ethical use, transparency, accountability, electoral integrity and public trust.
Liberia was represented by Commissioner Barsee Leo Kpangbai and S. Isaac Zahn, the commission's deputy executive director for information and communications technology. The workshop also brought together electoral commissioners, technology specialists and other regional stakeholders, including representatives associated with Senegal and Sierra Leone.
The case for regional standards is strong. Elections are national processes, but the platforms, vendors and information networks affecting them cross borders. A deceptive media technique or poorly governed data tool tested in one country can quickly appear in another. Shared guidelines can help commissions compare risks, establish common procurement expectations and exchange evidence about what works.
AI can support administrative tasks such as voter communication, document processing, translation, logistics and anomaly detection. Those uses may improve speed or access, but they also create risks if an automated system produces false information, treats groups unevenly or makes an important decision that officials cannot explain.
Public communication deserves its own rule set. Election commissions may use automated tools to answer routine questions, yet every response about registration, polling locations or results must be traceable to an authoritative record. Systems should clearly identify automated assistance, preserve official language versions and provide a rapid human escalation route when an answer could affect a citizen's ability to vote.
The guidelines should therefore distinguish low-risk assistance from uses that require stronger restrictions. Drafting a public information notice is not equivalent to determining voter eligibility. Systems that influence registration, accreditation, results management or enforcement need rigorous testing, human authority, documented appeals and independent scrutiny.
Nigeria's planned use of AI in preparations for its 2027 general election featured in the Abuja discussion. Africa AI Ledger has previously examined how Nigeria is preparing AI-related election defences, particularly around manipulated information and operational security. The regional process broadens that conversation from defensive tools to common rules for election administrators themselves.
Liberia is also considering how emerging technology could support its preparations for the 2029 general election. Its commission says AI tools are already being explored in communications and ICT functions. Before that experimentation expands, citizens should be told what systems are used, what data they access and which decisions remain exclusively human.
Validation is not the same as formal adoption or implementation. ECONEC and participating commissions should publish the final guidelines, explain how compliance will be assessed and create a mechanism for civil society, political parties and technical experts to challenge weaknesses before election day.
The region does not need AI in elections simply because the technology is available. It needs carefully governed tools that solve a defined administrative problem without making the electoral process harder for citizens to understand or trust.
Source & attribution
Source: necliberia.org