2027: Ex-INEC commissioner raises concerns over AI use in results management

A former National Commissioner of the Independent National Electoral Commission (INEC), Okechukwu Ibeanu, has warned that the commission’s plan to deploy Artificial Intelligence (AI) in election results management could undermine confidence in the 2027 elections if clear rules and safeguards are not established.

Mr Ibeanu said INEC should publish and implement an AI governance framework before using the technology in sensitive areas of election administration, particularly the processing and management of results.

He raised concerns about the proposed use of AI to extract figures from physical EC8A result sheets into electronic formats, saying Nigerians must know how the system would work, how discrepancies would be resolved and who would be accountable for decisions made through the technology.

Mr Ibeanu spoke at a plenary session titled “Governing AI in Nigeria’s 2027 Elections: Safeguards, Accountability and Elections Integrity” at the AI and Democracy Forum in Abuja.

The forum was organised by Yiaga Africa in partnership with INEC, Microsoft, TikTok, the Centre for Journalism Innovation and Development (CJID), Dataphyte, the MacArthur Foundation, the International Foundation for Electoral Systems (IFES), the German Agency for International Cooperation (GIZ) and other partners.

INEC has power to regulate AI use

Mr Ibeanu rejected suggestions that INEC had run out of time to develop regulations ahead of the 2027 elections.

He said the commission could immediately issue preliminary guidelines under its existing legal powers while working towards a more comprehensive framework.

“I don’t agree that it is late. It is possible for INEC to make a set of regulations, however preliminary they may be, to make sure that some of these things are brought under control,” he said.

He cited Section 151 of the Electoral Act, which empowers INEC to make regulations, guidelines and manuals for the administration of the law.

He also referred to Section 160 of the Constitution, which provides for the commission’s autonomy in regulating its internal affairs.

The former commissioner said INEC had already developed a two-year roadmap for AI adoption, with governance identified as one of its key components.

He said he was involved in the establishment of the commission’s AI division in 2025 and urged the current leadership to implement the roadmap while being more open about its plans.

“One of the main issues in that roadmap was the whole question of governance of AI. And one of the things that INEC itself said it was going to do was to create a committee that will oversee this,” he said.

According to him, the proposed committee was expected to include civil society representation.

He urged INEC to disclose the specific functions of the committee and explain how Nigerians would be able to scrutinise the commission’s use of AI.

Questions over electronic EC8A

Mr Ibeanu said the most immediate concern was the use of AI in election results management.

He said the proposed electronic EC8A system appeared designed to use AI to extract figures from physical EC8A result sheets, but questioned how the electronic version would relate to the original result sheet signed at the polling unit.

“It is not even clear exactly how this is going to work. If there are contradictions, for example, between the EC8A and the electronic EC8A, what will happen?” he asked.

He said INEC must clarify which document would take precedence where discrepancies occur and establish a transparent mechanism for resolving such disputes.

The former commissioner also questioned how figures extracted by AI would be transmitted, verified and processed at the back end.

He said Nigerians should not be expected to trust an electoral technology they cannot independently understand or scrutinise.

“The danger is to find external suppliers and hand over these things to them. And that will be very, very risky for managing elections,” he said.

He therefore urged INEC to build sufficient internal technical capacity and ensure that the procurement and deployment of AI were governed by clearly defined institutional rules.

‘Algorithmic war’ looms

Mr Ibeanu also warned that the use of AI in the 2027 elections would extend beyond INEC, with political parties and candidates likely to deploy technology to analyse election data and independently interpret results.

He described the potential contest as an “algorithmic war” in which political actors could increasingly rely on technology to collect, analyse and interpret election results.

He called for regulations governing political parties’ use of AI during campaigns, including requirements for labelling AI-generated campaign materials and sanctions for violations.

According to him, transparency must begin before election day, with INEC providing sufficient information about the technologies it intends to deploy and how they will affect the electoral process.

Yiaga demands publication of AI framework

Also speaking, Cynthia Mbamalu, director of Programmes at Yiaga Africa, called on INEC to immediately publish its AI governance framework so that political parties, civil society organisations, the media and other stakeholders can scrutinise it.

Ms Mbamalu said Nigerians needed to know how election results would be captured, transmitted, verified and collated under the proposed system.

She specifically called for the publication of a comprehensive results management protocol covering the process from polling units to the final collation stage.

“We want from INEC a protocol on results management that clearly states the process from the polling units,” she said.

She said the protocol should identify those responsible for verifying results, the criteria for flagging anomalies and the procedures for resolving discrepancies.

Ms Mbamalu also called for independent audits of INEC’s systems before the elections and urged the commission to involve civil society organisations, political parties, the media and other qualified stakeholders in the process.

She said INEC’s planned stress tests should extend beyond voter accreditation to cover the uploading and transmission of election results.

She also raised concerns about cyberattacks against electoral infrastructure and urged the government to ensure that election security was adequately addressed in any national AI governance legislation.

However, she cautioned against regulations being used to restrict freedom of expression, suppress dissent or discourage technological innovation.

She also urged technology companies to work with electoral stakeholders to tackle AI-generated misinformation and other threats that fall outside INEC’s direct control.

IFES warns against blind trust in vendors

Lisa Poggiali, Chief AI Adviser at IFES, said AI governance in elections required collaboration between election management bodies, technology experts, policymakers and institutions responsible for oversight.

She said while principles such as transparency, inclusion, accessibility, integrity and respect for human rights were widely accepted, the challenge was translating those principles into practical decisions about procurement and deployment.

Ms Poggiali warned that technology vendors could make claims about the accuracy and reliability of their AI systems that election officials might not be able to independently verify.

She called for transparency in procurement, independent audits of AI models and documented procedures for dealing with anomalies identified by automated systems.

She said human officials must retain responsibility for decisions involving AI, with identifiable personnel assigned to review anomalies and document how they were handled.

READ ALSO: INEC explains exclusion of Pantami, Adebutu from 2027 governorship candidates’ list

Ms Poggiali also cautioned that different AI models could produce different results because of variations in their training data and algorithms.

She said election administrators should therefore understand how models were developed, the data used to train them and how their outputs were generated.

“AI is not something that we should be trying to deploy in every single aspect of the election,” she said.

She said electoral authorities should first identify the problem they were trying to solve and determine whether AI was actually the appropriate solution.

According to her, some technologically advanced election management bodies had deliberately chosen not to use AI in certain areas because of concerns about public trust.

Deepfakes, misinformation threaten turnout

Sami Suleiman of Paradigm Initiative highlighted the risks posed by AI-generated disinformation, deepfakes, manipulated election results and weaknesses in social media content moderation.

He warned that old videos of attacks could be recirculated as though they were current incidents, potentially frightening voters and discouraging them from participating in elections.

He also raised concerns about genuine evidence being dismissed as AI-generated, inadequate moderation of African languages and inconsistent enforcement of online regulations.

Mr Suleiman advocated a risk-based and human-centred approach to AI governance, including independent electoral oversight, rapid-response mechanisms, independent audits, enforcement procedures and avenues for appeal.

He said some AI systems deployed in Africa were developed using datasets that might not adequately represent local populations, creating risks of discrimination and unreliable outcomes.