By Simpson Global Media News Desk
AI Moves From Policy to Operational Use
Nigeria’s Independent National Electoral Commission is expanding its use of artificial intelligence in election technology, with automated AI auditing tools being integrated into the verification of election results and AI-driven systems being deployed to identify anomalies and protect the commission’s digital infrastructure ahead of the 2027 general elections.
INEC Chairman, Professor Joash Amupitan, disclosed the development on Tuesday, October 6, 2026, during a World Press Conference and National Stakeholders’ Meeting in Abuja marking 100 days to the first ballot.
The commission says the technology will operate under a five-pillar AI governance framework with mandatory human oversight. It is also finalising an AI governance policy to guide how artificial intelligence will be used within the electoral process.
The announcement represents a further stage in INEC’s effort to incorporate artificial intelligence into election administration.
The commission had already announced the creation of an Artificial Intelligence Division within its Information and Communications Technology Department in May 2025. At that time, INEC said the division would coordinate the use of AI while addressing both the opportunities and risks created by the technology.
The latest development is therefore less about the creation of an entirely new idea and more about moving that institutional framework towards practical deployment in the 2027 election cycle.
What the New AI Systems Are Expected to Do
According to Amupitan, automated AI auditing tools are being integrated into the commission’s results-verification process.
The objective is to use technology to identify discrepancies or unusual patterns that may require further examination by election officials.
INEC has also said it is working with international technology partners to integrate AI-driven anomaly-detection tools capable of monitoring database traffic and helping shield its public portals against cyber intrusions or unauthorised alterations.
That distinction is important.
The commission is not saying that an artificial intelligence system will independently determine which candidate has won an election.
Instead, AI is being positioned as an additional layer of technological analysis around processes that remain under the control of human officials.
Amupitan specifically said the automated checks would operate under mandatory human oversight.
In practical terms, an AI system could flag a result or data pattern for examination, but the final determination would remain subject to established electoral procedures and human verification.
That approach is significant because election technology must balance automation with accountability.
Why Anomaly Detection Matters
Election results generate enormous quantities of data.
At polling-unit level, figures are recorded, transmitted, stored and eventually aggregated through several stages of the electoral process.
Any technology capable of rapidly comparing information across those stages could potentially help officials identify inconsistencies more quickly.
An anomaly does not automatically mean electoral fraud.
A difference in data may result from a legitimate administrative problem, a transmission error, an input mistake, a network interruption or another technical issue.
That is why human review remains essential.
An AI system can identify something unusual. It cannot, by itself, provide the legal or administrative context necessary to determine why the discrepancy occurred.
This is one reason INEC’s decision to place the technology under a governance framework is important.
The credibility of an AI-assisted electoral system will depend not simply on whether algorithms can identify unusual patterns, but also on whether officials can explain how those alerts are investigated and resolved.
The Technology Challenge Is Bigger Than Counting Votes
INEC’s latest technology plans extend beyond results verification.
The commission is also dealing with cybersecurity, database protection, digital result transmission, biometric accreditation and the security of its public-facing online systems.
Amupitan said AI-driven anomaly detection tools would monitor database traffic and help protect INEC’s public portals against cyber intrusions or unauthorised alterations.
That comes as Nigeria’s cybersecurity authorities are warning that attackers are increasingly using artificial intelligence and other automated technologies.
The Nigeria Computer Emergency Response Team, ngCERT, issued several security advisories on October 6, including warnings involving malware campaigns, vulnerabilities and emerging AI-enabled threats. It has also warned organisations that AI agents can increasingly be used to automate sophisticated cyberattacks.
The Nigerian Communications Commission’s Computer Security Incident Response Team has separately warned of AI-driven social engineering and deepfake threats, including synthetic voice calls, phishing messages and impersonation scams.
For an institution responsible for managing election data, these developments create a dual technology challenge.
INEC must use increasingly sophisticated digital tools while simultaneously defending the systems on which those tools depend.
Deepfakes Create a New Election Problem
Another concern identified by INEC is the growth of generative artificial intelligence.
Amupitan warned that generative AI, voice cloning and deepfake video technology could make it easier to fabricate the voices of election officials, manipulate result sheets and create counterfeit official documents.
This creates a problem that differs from traditional cyberattacks.
A cyberattack may attempt to compromise a database or disrupt a website.
A deepfake may instead attempt to compromise public confidence.
A fabricated audio recording appearing to feature an election official could circulate through WhatsApp, Facebook, TikTok or other platforms within minutes.
A manipulated image of a result sheet could be presented as evidence of an electoral development that never happened.
A fake video could appear to show an official announcing a result or making a controversial statement.
The technology required to create convincing synthetic material is becoming increasingly accessible.
That means the challenge for election administrators is no longer confined to securing servers and networks.
They must also help citizens, journalists and political stakeholders distinguish authentic information from convincing fabrications.
The Threat to Public Trust
The most significant consequence of AI-generated election misinformation may not be a single false claim.
It may be the cumulative effect of repeated falsehoods.
If citizens encounter enough fabricated documents, fake videos and synthetic audio recordings, they may begin to doubt genuine information as well.
That creates what Amupitan described as a broader threat to public confidence in legitimate information from constitutional authorities.
The problem is particularly serious during elections because information often moves faster than formal verification.
A false claim about a polling unit can be shared thousands of times before officials have an opportunity to investigate it.
By the time a correction is published, the original material may already have reached a much larger audience.
Artificial intelligence therefore creates a race between fabrication and verification.
Election institutions need technology capable of responding quickly without sacrificing accuracy.
Human Oversight Remains Central
The decision to retain mandatory human oversight is therefore one of the most important parts of INEC’s approach.
AI can process large volumes of information rapidly, but election decisions carry legal, political and constitutional consequences.
An automated system should not be allowed to turn an unexplained statistical anomaly into an allegation of wrongdoing.
Similarly, a technical flag should not automatically invalidate a result.
Human officials must investigate the underlying circumstances.
This is particularly important because AI systems can themselves make mistakes.
Algorithms are trained and configured by people. Their effectiveness depends on the quality of data available to them, the assumptions built into their design and the circumstances in which they are deployed.
A system designed to detect unusual election results must therefore be tested extensively before being trusted in a live electoral environment.
Nigeria Has Already Been Building Its Digital Election Architecture
The move towards AI is taking place on top of several years of investment in digital election technology.
Nigeria has progressively introduced biometric accreditation and electronic result-transmission systems as part of efforts to improve election administration.
The Bimodal Voter Accreditation System, commonly known as BVAS, is one of the most visible examples.
The technology is designed to authenticate voters biometrically and support the accreditation process.
INEC has now said that its BVAS devices have undergone software hardening, battery optimisation and biometric calibration, while 57,000 new devices have been acquired for the 2027 elections.
The commission has also said the enhanced eEC8A system will digitise polling-unit result forms and facilitate their upload to the INEC Result Viewing Portal, known as IReV.
These technologies form part of a broader digital chain.
The voter must be properly registered.
The voter must be accredited.
The polling-unit result must be correctly recorded.
The result must be transmitted or uploaded through the appropriate system.
The information must be protected from unauthorised alteration.
And the final result must be verified and declared through the legally established process.
AI is being introduced into this wider architecture rather than operating as a stand-alone election technology.
More Devices Mean More Data and More Security Responsibilities
INEC’s acquisition of 57,000 additional BVAS devices will also increase the technological footprint of the 2027 election.
More devices mean greater capacity, but they also create additional requirements for logistics, maintenance, software management, authentication and cybersecurity.
Every device used in an election environment must be properly configured and accounted for.
Software updates must be controlled.
Authentication credentials must be protected.
Connectivity must be reliable where systems depend on networks.
Equipment must be secured during transportation and storage.
Officials using the devices must be adequately trained.
These are not purely technical questions.
They become election-integrity questions because a failure at any stage can affect the public’s confidence in the process.
AI Cannot Compensate for Weak Infrastructure
Another issue is the underlying infrastructure on which advanced election technology depends.
AI systems require computing capacity, secure networks, electricity, data storage and reliable communications.
A sophisticated algorithm cannot compensate for a broken network connection at a critical moment.
Likewise, a strong cybersecurity system cannot entirely eliminate the risks created by weak passwords, poor access controls, inadequate staff training or compromised devices.
Nigeria’s broader digital infrastructure has improved considerably, but uneven connectivity and power reliability remain important challenges.
This means that technological planning for elections has to include contingency arrangements.
Where connectivity fails, there must be clearly defined procedures.
Where a device malfunctions, replacement mechanisms must be available.
Where suspicious activity is detected, there must be a documented response.
And where an AI system produces an alert, trained personnel must be able to investigate it.
Governance Is as Important as Technology
INEC’s planned AI governance policy will therefore be closely watched.
The effectiveness of AI in election administration will depend not only on the quality of the software but also on the rules governing its use.
Questions that a governance framework should answer include who can access the system, what data it can process, how its decisions or alerts are documented, who can override automated recommendations, how errors are corrected and how citizens can challenge decisions affected by technology.
Transparency will also matter.
Election stakeholders are likely to want assurances that AI systems are not being used to secretly favour one outcome over another.
The best way to address such concerns is through clear rules, independent oversight, testing and documentation.
The commission has already emphasised that its AI deployment will operate under mandatory human oversight.
The next challenge will be demonstrating how that principle works in practice.
Testing Will Be Crucial
A live general election is not the appropriate environment for testing an unproven technological system for the first time.
INEC will therefore need to subject its AI tools and related digital infrastructure to extensive testing before January.
Testing should include normal election conditions as well as failure scenarios.
Officials need to know how the system behaves when data is incomplete.
They need to understand how it handles duplicate records or inconsistent entries.
They need to test how quickly it can identify unusual traffic patterns.
Cybersecurity teams need to simulate attacks.
Election officials need to practise responding to alerts.
The system must also be evaluated for false positives and false negatives.
A false positive occurs when a system identifies something as suspicious even though it is legitimate.
A false negative occurs when the system fails to identify a genuine problem.
Both can have consequences.
Too many false alerts could overwhelm election officials.
Too few could create a false sense of security.
AI and the Human Element
Technology does not eliminate the need for trained election officials.
Instead, it changes the nature of some of their responsibilities.
Officials may increasingly need to understand digital audit reports, cybersecurity alerts, data-quality indicators and automated anomaly assessments.
Training will therefore have to go beyond operating election equipment.
Personnel must understand what the technology can and cannot do.
They must know when an automated warning requires escalation.
They must also understand the importance of preserving evidence and documenting decisions.
This becomes particularly important if an election result is later challenged.
A credible digital audit trail can help demonstrate how a discrepancy was detected, who reviewed it and what action was taken.
The Media Will Become Part of the Verification Chain
INEC is also calling on media organisations to strengthen their fact-checking capacity.
Amupitan has urged editors to establish specialised fact-checking desks, use forensic image-analysis tools and verify election information with INEC before publication.
This reflects a broader change in the information environment.
Traditionally, election reporting depended heavily on physical documents, statements from officials and direct observation.
Today, journalists must also assess digital evidence.
That includes photographs, videos, screenshots, audio recordings, social-media posts and documents circulated electronically.
AI-generated material makes that task more complicated.
A newsroom may receive an apparently authentic recording of an official before the official has even made a public statement.
Journalists will increasingly need tools and procedures for verifying the origin and authenticity of such material.
Citizens Will Also Need Digital Literacy
The responsibility cannot rest entirely with INEC or the media.
Voters themselves will be confronted with increasingly convincing synthetic content.
A short video showing an alleged result announcement may look authentic.
A voice message may sound exactly like a familiar official.
A photograph of a result sheet may appear convincing.
Citizens therefore need to develop simple verification habits.
Before forwarding election-related information, users should consider where it came from, whether the source is identifiable and whether credible official channels have confirmed it.
A message that creates extreme urgency or encourages users to spread information immediately should receive additional scrutiny.
The goal is not to make citizens suspicious of everything they see online.
It is to prevent speed from replacing verification.
The Cybersecurity Environment Is Changing
INEC’s technology strategy also comes at a time when cybersecurity itself is changing rapidly.
Nigeria’s ngCERT warned on October 5 that AI agents are being used to automate parts of sophisticated cyberattack chains, including vulnerability exploitation and malware development.
On October 6, the organisation listed additional advisories involving threats affecting software, network infrastructure and endpoint systems.
The NCC-CSIRT has similarly warned that criminals are using AI to produce increasingly convincing phishing messages, synthetic voices and impersonation attempts.
For election administrators, the implication is straightforward.
Cybersecurity cannot be treated as a one-time exercise completed before election day.
It must operate continuously.
Systems need monitoring before, during and after the election.
Suspicious access attempts need to be investigated.
Software vulnerabilities need to be patched.
Personnel need to understand social-engineering risks.
And incident-response teams need to be prepared for situations in which attackers attempt to disrupt public-facing systems.
Protecting the Public Results Portal
One of INEC’s stated objectives is to protect its public portals from cyber intrusions and unauthorised alterations.
That is important because online publication of election information makes the integrity of public-facing systems part of the wider credibility of the election.
If a legitimate portal becomes unavailable, citizens may turn to unofficial sources.
If false information appears to come from an official platform, the damage could be greater.
If attackers alter or manipulate information, even temporarily, they could create confusion.
The commission’s plan to monitor database traffic using AI-driven anomaly detection is therefore intended to provide another layer of protection.
But again, technology must operate alongside conventional cybersecurity controls.
Strong access management, encryption, secure software development, network segmentation, backups and trained security personnel remain necessary.
AI is an additional capability, not a substitute for basic cybersecurity.
The International Dimension
INEC says it is working with international technology leaders and partners as it develops its AI-driven anomaly-detection capabilities.
Such partnerships can provide access to specialised expertise and technology.
They also raise questions about data governance and technological dependence.
Election data is highly sensitive.
Any external technology partner involved in election systems must operate within clearly defined legal, security and privacy requirements.
Nigeria will need to ensure that critical election infrastructure remains under appropriate national control and that contractual arrangements provide sufficient protection for sensitive information.
The technology partnership question is therefore not simply about acquiring the most advanced tool.
It is about building a system that Nigeria can securely operate, audit and maintain.
AI Does Not Replace Electoral Law
Another important boundary is the relationship between technology and law.
Artificial intelligence can identify patterns.
It cannot create legal authority.
The Electoral Act and established electoral procedures determine how votes are cast, counted, transmitted, collated and declared.
If an AI system identifies a discrepancy, officials must still follow the appropriate legal process for investigating it.
Likewise, an automated system cannot independently overturn a polling-unit result or determine that a candidate has won.
This legal boundary is essential to maintaining accountability.
Technology should strengthen established procedures rather than create an alternative decision-making system outside them.
Building Confidence Through Explainability
For AI-assisted election technology to gain public confidence, INEC may also need to explain its operation in language that ordinary Nigerians can understand.
Technical descriptions of algorithms are unlikely to be sufficient.
Citizens will want to know what the system does, what it does not do and who makes the final decision.
For example, if an AI tool flags an unusual result, the public should understand that the alert is an indication for human review rather than proof of wrongdoing.
Similarly, if a cybersecurity system detects suspicious database traffic, officials should have procedures for investigating the incident before conclusions are announced.
Clear communication can prevent the technology itself from becoming a source of confusion.
The 100-Day Countdown Adds Urgency
The announcement comes as Nigeria enters the final 100-day phase before the first 2027 ballot.
INEC says it has completed nine of 14 major milestones in its election timetable and has moved from planning towards execution.
The presidential and National Assembly elections are scheduled for January 16, 2027, while governorship and State Assembly elections are scheduled for February 6, according to INEC's election preparations.
That leaves limited time for technology testing, training and public communication.
The pressure is therefore not only to introduce new systems but to demonstrate that they are ready.
A Larger Transformation of Election Technology
Nigeria’s election technology journey is moving into a new phase.
The first major challenge was digitising processes that had historically depended heavily on paper and manual procedures.
The next challenge is making digital systems more intelligent, resilient and secure.
AI represents part of that transition.
Instead of simply recording information, systems can increasingly analyse large datasets, identify unusual patterns and alert human officials to possible problems.
That capability can be useful if deployed responsibly.
But the technology also introduces new risks.
AI can help an election commission detect manipulation, but it can also help criminals manufacture convincing false information.
It can strengthen cybersecurity, but attackers can also use AI to automate attacks.
It can improve data analysis, but poor data or poorly designed algorithms can produce misleading conclusions.
The technology is therefore neither inherently beneficial nor inherently dangerous.
Its value depends on how it is designed, governed and used.
What Nigerians Should Watch For
As the 2027 elections approach, several technology-related developments will be worth watching.
The first is the completion of INEC’s AI governance policy.
The second is the testing of its automated result-auditing tools.
The third is how the commission handles AI-generated alerts and whether procedures for human review are clearly defined.
The fourth is the cybersecurity performance of the commission’s public portals and databases.
The fifth is the readiness of the 57,000 newly acquired BVAS devices and their integration into the wider election system.
The sixth is the security of result transmission and publication.
And the seventh is the ability of INEC and other stakeholders to respond quickly to deepfakes and synthetic misinformation.
Together, these factors will determine whether the technology strengthens confidence or creates additional uncertainty.
The Importance of Independent Scrutiny
The introduction of AI should also attract scrutiny from technology experts, civil society organisations, election observers and other stakeholders.
Independent examination can help identify weaknesses before election day.
Questions about privacy, cybersecurity, algorithmic bias, auditability and transparency should be addressed early.
Testing results should, where appropriate, be communicated to the public.
The more important the technology becomes, the more important independent oversight becomes.
An election system should not depend entirely on public trust in the institution operating it.
It should also provide verifiable processes that allow stakeholders to assess whether the system worked as intended.
Technology Must Serve the Voter
Ultimately, the purpose of election technology is not to make elections appear more sophisticated.
It is to make them more reliable, transparent, accessible and secure.
For voters, that means being able to register and authenticate successfully, cast a ballot without unnecessary technological barriers, and have the resulting votes handled through processes that can be independently understood and scrutinised.
AI will be useful if it contributes to those objectives.
It will be counterproductive if its complexity makes the electoral system harder for citizens to understand.
The emphasis should therefore remain on reliability and accountability rather than technological novelty.
Nigeria’s AI Experiment Enters a Critical Stage
INEC’s latest announcement marks an important moment in Nigeria’s technology story.
The country is not merely discussing artificial intelligence as a future possibility.
It is beginning to integrate AI into one of the most sensitive public-sector processes: the management and verification of election information.
The commission says automated AI auditing tools are already being integrated into result verification, while AI-driven anomaly-detection systems are being developed to monitor database traffic and protect public portals.
At the same time, Nigeria’s cybersecurity agencies are warning that artificial intelligence is being used by attackers to increase the scale and sophistication of cyber threats.
That creates a technological race in which defenders and attackers are both gaining new capabilities.
INEC’s decision to retain mandatory human oversight is therefore critical.
Artificial intelligence should assist trained officials, not replace institutional responsibility.
The success of the system will ultimately depend on whether the technology is tested thoroughly, governed transparently, protected against attack and integrated with established electoral procedures.
The commission’s earlier decision in 2025 to establish an AI division provided the institutional foundation. The current deployment of AI auditing and anomaly-detection tools takes that strategy further towards operational use.
The next few months will determine how effectively that transition is managed.
Nigeria now has an opportunity to demonstrate that advanced technology can strengthen public institutions without weakening human accountability.
For the 2027 elections, the central question will not simply be whether INEC uses artificial intelligence.
It will be whether Nigerians can see, understand and verify how that technology contributes to a credible electoral process.
In an environment where a genuine document and a sophisticated fake can increasingly look alike, that distinction could become one of the defining technology challenges of Nigeria’s next election cycle.
Comments
Post a Comment