2027 Elections: Amupitan Must Ensure That Algorithms, Human Manipulation And Institutional Failures Do Not Rob Nigerians of Their Votes

The declaration by Professor Joash Amupitan, Chairman of the Independent National Electoral Commission (INEC), that algorithms must never determine the outcome of Nigeria’s 2027 general elections is an important affirmation of democratic principles. Yet, beyond the reassuring sound of his words lies a more complicated question: how can INEC ensure that technology serves the electorate rather than undermines its confidence in the electoral process, particularly in a country where disputes over election administration, result transmission and institutional impartiality have repeatedly tested public faith in democracy?

Speaking at the Artificial Intelligence and Democracy Forum 2026 in Abuja, Amupitan insisted that the decision about who governs Nigeria belongs exclusively to Nigerian voters. He acknowledged that artificial intelligence would influence the information citizens receive and the organisation of elections but maintained that human oversight must remain in critical decisions concerning voter accreditation, result collation and candidates’ qualifications. He also disclosed that INEC had completed 10 of the 14 major operational milestones under the Electoral Act 2026, while introducing online voter registration and digitised candidate nominations.

These are consequential assurances at a time when artificial intelligence is transforming political communication, information management and electoral administration around the world. However, the credibility of the 2027 elections will not be secured by declarations alone. It will depend on whether INEC translates its stated principles into enforceable safeguards, transparent procedures and demonstrable accountability before, during and after voting.

The Principle Is Right, But The Challenge Is Bigger

At its core, Amupitan’s declaration expresses a fundamental democratic truth: elections exist to translate the freely expressed choices of citizens into legitimate political authority. An algorithm, however sophisticated, cannot possess the democratic mandate to select a president, governor, senator or member of a legislative assembly. Technology may assist the process, but it cannot replace the sovereign will of the electorate.

An algorithm is essentially a set of computational instructions used to process information and produce outputs. Artificial intelligence systems can analyse enormous volumes of data, identify patterns, generate content and support administrative decisions at speeds that would be difficult for human operators to match. Properly designed and deployed, such capabilities could improve voter services, detect irregularities, support logistics and make election administration more efficient.

The danger emerges when technological systems become opaque, inadequately supervised or vulnerable to manipulation. An algorithm trained on incomplete or inaccurate information can produce misleading results. An automated system may reject legitimate information, flag genuine voters incorrectly or reproduce biases embedded in its design. Where a system influences politically consequential decisions without a transparent explanation or an effective avenue of appeal, technology can become a source of disenfranchisement rather than a tool of democratic progress.

Nevertheless, it is important to distinguish between an algorithm that supports electoral administration and one that actually determines an election result. Software used to organise registration records, detect duplicate entries or assist with logistics does not necessarily decide who wins. The decisive questions are what authority has been delegated to the system, how its output is verified, who controls it and whether independent observers can establish that its operation has not compromised the integrity of the vote.

Amupitan is therefore right to reject the prospect of machines replacing the electorate. But the commission must also explain precisely how it will prevent technological systems from influencing election outcomes indirectly through flawed accreditation, inaccurate records, manipulated information or unreliable result-management processes.

Human Oversight Must Not Become A Cover For Human Manipulation

One of the most significant aspects of Amupitan’s statement is his insistence that human oversight will remain in place for voter accreditation, result collation and candidate qualification. This is reassuring in principle, but it raises a difficult question: who supervises the human beings exercising that authority, and what happens when their decisions are challenged?

The assumption that human involvement automatically guarantees fairness is questionable. Nigeria’s electoral challenges have never been confined to technology. Human beings can make honest mistakes, but they can also act under political pressure, compromise established procedures, manipulate records or deliberately obstruct the implementation of electoral safeguards. A system that retains human control without effective accountability may simply transfer the possibility of manipulation from a machine to an official.

Indeed, the most dangerous electoral arrangement would be one in which technology is blamed whenever a discrepancy occurs, while officials retain discretionary powers that are difficult for voters, political parties and independent observers to scrutinise.

INEC must therefore establish a clear distinction between human supervision and unchecked human discretion. Decisions should be governed by published rules, recorded in auditable systems and open to timely review. Where an automated process flags a voter, rejects a nomination or identifies an apparent discrepancy, the affected individual or political party should have access to a clear explanation and an effective procedure for seeking correction.

Similarly, where an electoral official overrides a system-generated recommendation, the action should be recorded, justified and subject to appropriate review. Such safeguards would protect the commission against both technological errors and the possibility that individuals could exploit their authority to manipulate electoral administration.

Human judgment is indispensable in a democracy, but it must be exercised within the law. The objective should not be to choose between machines and people as though either is inherently trustworthy. It should be to construct a system in which technology, human operators and independent verification reinforce one another, with no individual or institution enjoying unchecked power over the electorate’s choice.

Artificial Intelligence Could Distort The Election Before Voting Begins

Amupitan’s acknowledgement that artificial intelligence will influence the information voters receive is particularly important. The threat posed by AI extends well beyond the physical polling station and the electronic systems used to manage election results.

Generative AI can produce convincing but fabricated photographs, audio recordings, videos and statements attributed to political candidates. A manipulated recording appearing to show a presidential candidate making an inflammatory remark could circulate across social media before the candidate has an opportunity to respond. Fabricated announcements about changes to polling arrangements, false claims about a candidate’s disqualification or misleading reports of electoral violence could discourage voters from participating.

The consequences could be serious in a politically competitive environment where public emotions run high and information spreads rapidly through messaging applications and social networks. By the time a false claim is disproved, it may already have influenced voter perceptions, intensified ethnic or religious tensions or undermined confidence in legitimate electoral procedures.

There is also the danger of algorithmic amplification. Social media platforms frequently use recommendation systems to determine which content users encounter. Material that attracts attention, generates outrage or provokes arguments can spread rapidly, irrespective of its accuracy. Political actors could exploit these mechanisms to target voters with contradictory messages, deepen existing divisions or overwhelm the public sphere with misleading information.

This is where Amupitan’s warning must move beyond the question of whether an algorithm can determine the final vote count. A voter whose choice has been shaped by a fabricated recording, systematic disinformation or deceptive political advertising may still cast a ballot personally, but the quality of the information informing that decision has been compromised.

Democracy does not require voters to agree on political questions. It requires that they have a meaningful opportunity to make their choices without coercion, intimidation or deliberate deception that substantially undermines informed participation.

INEC cannot regulate every online conversation, nor should it assume powers that properly belong to other institutions. However, it can work with relevant agencies, civil society organisations, media houses, technology platforms and political parties to develop rapid verification mechanisms, public education campaigns and clear procedures for addressing election-related disinformation. Such cooperation must respect freedom of expression and avoid turning the fight against falsehood into a pretext for suppressing legitimate criticism of the commission or the government.

The credibility of this effort will depend on impartiality. False information circulated against the ruling party must be treated according to the same standards as false information circulated against opposition parties. INEC must not become an arbiter of political opinion; its responsibility is to protect the integrity of the electoral process and help citizens distinguish official electoral information from demonstrably fabricated material.