INEC Boss, Prof Amupitan, Says Artificial Intelligence Is Not An Enemy of Democracy

Being Remarks by the Chairman, Independent National Electoral Commission (INEC), Professor Joash O. Amupitan, SAN, at the Opening Session of the National Summit on Electoral Information Integrity on Monday, 2^nd November 2026, in Abuja

Introduction: It is a great pleasure to stand before you this morning at the opening session of this National Summit on Electoral Information Integrity, themed “Beyond the Ballot: AI, FIMI, and the Integrity of Nigeria’s 2027 Elections.”

As we convene here at the Continental Hotel, we are exactly where we need to be just a few weeks ahead of the next general election, proactively addressing the most significant technological challenge to modern democracy- the weaponisation of information.

Therefore, this summit comes at a pivotal moment as our nation actively counts down to 16th January 2027, when the first ballots will be cast for the Presidential and National Assembly elections. Indeed, the urgency of safeguarding our information space was so compelling that we rescheduled the South-South leg of our ongoing zonal readiness assessment and staff engagements in Port Harcourt to participate directly in this crucial dialogue.

The Independent National Electoral Commission (INEC) views this summit, supported under the EU-SDGN III framework, not just as a conference, but as a critical defence council for the integrity of our democratic process.

It was John F. Kennedy in his 1960 presidential campaign speech that emphasised that voter ignorance threatens collective security. The same may also be said of electoral information disorder. It is also a threat to our collective security.

2. The Changing Landscape of Electoral Vulnerability

For decades, those who manage elections in emerging democracies like ours focused their primary energy on physical and operational vulnerabilities. They secure sensitive paper ballots, ensure early logistics transport across difficult terrains, maintain strict chains of custody, and protect electoral officials at remote voting locations. While those operational realities remain essential, the terrain of electoral integrity has fundamentally shifted.

Today, elections can be undermined long before the first ballot is cast. The threat is no longer limited to physical interference (such as violent ballot snatching) at the polling unit; it takes shape within digital networks through synthetic media, automated misinformation/disinformation, deepfake fabrications, digital micro-targeting, and Foreign Information Manipulation and Interference (FIMI). The troubling reality is that many democracies are not really prepared for these.

As we prepare for the 2027 General Election, the undeniable truth that we must confront is that public awareness alone is no longer an adequate defence against synthetic content and digital manipulation. Informing citizens that deepfakes exist or urging caution when consuming online news is useful, but it is insufficient against well-coordinated, highly funded, and hyper-targeted information operations designed to erode public trust in election administration.

What Nigeria urgently requires is a trusted, coordinated, and rights-respecting operational ecosystem—an institutional structure through which public authorities, security bodies, independent fact-checkers, media networks, and platform operators can collectively anticipate emerging threats, verify evidence, escalate incidents without delay, communicate authoritatively, and institutionalise post-event lessons.

That is why INEC welcomes this gathering organised under the EU-SDGN framework. The objectives set for this summit- from mapping out high-risk moments across the electoral cycle to adopting a concrete Commitment on Electoral Information Integrity— address the real-world operational challenges we face on the ground.

3. Understanding the AI and FIMI Threat Landscape in 2027

Artificial Intelligence is not inherently an enemy of democracy. Within election management bodies worldwide, machine learning assists in deduplicating large voter registries, refining logistics planning, streamlining data analytics, and improving administrative efficiency.

However, the rapid democratisation of generative AI models has dramatically lowered the barrier to entry for malicious actors. Capabilities that once required state-level intelligence resources- such as high-fidelity voice cloning of public figures, hyper-realistic video synthesis of election officials making false announcements, or automated bot networks simulating organic public outrage- are now accessible at negligible cost.

During an election cycle, these tools target specific institutional vulnerabilities across distinct timeline windows. Let’s examine the following examples:

Pre-Election Operations: The generation of fake operational directives, forged registration portals, or manipulated claims regarding voters’ list updates aimed at causing confusion during the continuous voter registration and display exercises.

Campaign and Eve-of-Election Windows: Section 98 (1) of the Electoral Act, 2026 provides that political campaigns shall end 24 hours prior to polling day. For the Presidential & National Assembly polls, campaigns will end on 14th January, 2027 and 4th February, 2027 for the Governorship and the State Houses of Assembly. But deepfake audio or video drops can be deliberately timed and specifically calculated to exploit moments like this when media verification channels are constrained and candidate response windows are narrow. With the opening of Campaigns in August 2026 for the 2027 elections, political parties and their supporters have engaged social media to promote their candidates and denigrate their opponents through false information.

A clear global precedent occurred during the 2023 Slovak Parliamentary Election, demonstrating how passive public awareness fails when synthetic media exploits high-vulnerability operational windows.

Just two days before Slovak voters went to the polls- during a mandatory 48-hour electoral blackout period when political parties and traditional news outlets were legally restricted from campaigning or broadcasting political responses—an AI-generated audio recording went viral on social media. The cloned audio purported to capture a high-ranking political party leader actively discussing a scheme to buy votes and manipulate election results.

Because the AI-generated audio drop was deliberately timed to coincide with the blackout period when campaigning was statutorily prohibited, fact-checkers, election authorities, and media houses could not issue timely debunking broadcasts before millions of voters hit the polls. Even though the electorate had been broadly informed that AI deepfakes were a rising global trend, the sheer speed of digital dissemination combined with restricted response windows paralysed institutional verification efforts.

This case proves that public awareness alone cannot counter well-funded, hyper-targeted digital manipulation. Therefore, protecting our polls requires rapid-response escalation protocols capable of verifying facts and counteracting viral deception in real time.

Collation and Result Announcement: Synthetic audio or manipulated documentation falsely announcing declared results or declaring cancelled polls in key battleground constituencies to trigger civil unrest and undermine formal collation processes.

When local political actors leverage generative AI to manipulate domestic perceptions, or when foreign entities execute coordinated FIMI campaigns to destabilise national stability, the casualty is not merely candidate reputation- it is citizen trust in the democratic process itself.

4. INEC’s Readiness Priorities

Our commitment to 2027 is rooted in technological resilience and transparency. We are prioritising:

Voter Education. Moving beyond “how to vote” to “how to verify,” ensuring citizens can distinguish official INEC communications from AI-generated deepfakes. Educating citizens on genuine platforms or Apps they can use to quickly verify deepfakes. As Aminu Kajuru, a digital communication expert, puts it, Nigeria should not wait until false information goes viral before responding, but institutions must use “pre-bunking,” method by educating citizens in advance about common misinformation tactics and giving them simple tools to verify information and, in addition, use short audio messages in local languages and Pidgin English, particularly through WhatsApp, to teach citizens how to identify fabricated information.

Technological Safeguards: Strengthening our digital infrastructure to withstand Foreign Information Manipulation and Interference (FIMI). This requires investments in infrastructure and the support of our vendors.
Real-time Response: Establishing robust internal mechanisms to debunk viral misinformation within minutes of its emergence.

Partnering with global social media platforms such as TikTok, Facebook, Instagram, WhatsApp, YouTube, Facebook Messenger, Telegram, X (formerly Twitter), etc. We already have an MOU with TikTok.

5. Institutional Concerns

A common mistake made by institutions when responding to modern disinformation is relying entirely on reactive debunking. Fact-checking is indispensable, but when a manipulated video achieves viral saturation within thirty minutes of release, a formal text rebuttal issued three hours later struggles to undo the cognitive damage.

We must be candid. Our primary concern is that while technology evolves at the speed of light, the truth often travels at the speed of trust. AI-enabled manipulation- from synthetic audio of election officials to “deepfake” videos of candidates- can trigger real-world disenfranchisement and civil unrest. The challenge is not just the “fake news” itself, but the “Liar’s Dividend,” where the mere existence of AI makes voters doubtful of what is actually true.

The Independent National Electoral Commission recognises that safeguarding the 2027 polls requires an institutional approach such as detection, verification, escalation, communication/ public messaging and post-incident learning.

6. Expectations of a Multi-Stakeholder Response

INEC cannot police the truth alone. Our expectation for this summit is clear: we need a United Front for Information Integrity.

First, our internal communications, ICT, and voter education structures are evolving. We are upgrading our technical capacity to monitor public data streams for early signs of coordinated manipulation targeting election operations. We are enhancing our capacity and workflows to detect deepfakes, verify suspect media, and issue authoritative clarifications directly through established digital and traditional broadcast channels. However, we cannot do this alone; hence this gathering.

Second, technical preparedness within the election management body alone is insufficient. When an AI-generated deepfake imitates an election official declaring a postponed poll, the response cannot depend on isolated press releases. It requires an immediate, pre-coordinated response protocol:

Platform operators must have fast-track escalation pathways to review and tag or remove demonstrably altered content that violates platform standards regarding electoral interference.

Telecommunications providers and media organisations must maintain clear channels to broadcast verified institutional alerts without administrative delays.

Fact-checkers and independent journalists must have direct access to designated INEC Communication desks to confirm operational facts within minutes rather than hours.

Security agencies must distinguish between protected political expression and actionable digital crimes aimed at destabilising public order.

Therefore:

To the Media: Be the first line of verification. Speed is important, but accuracy is the currency of democracy.

To Civil Society: Help us bridge the digital literacy gap in rural communities where AI manipulation can be most damaging.

To Technology Partners: We need tools that can detect and flag harmful synthetic content specific to the Nigerian linguistic and social context.

7. Balancing Security with Rights Safeguards

As we construct this information defense system, we must proceed with strict institutional restraint. The legitimate threat of AI manipulation must never become a pretext for state overreach, digital censorship, or the suppression of political dissent.

A free press, robust civil society critique, and open public debate are the lifeblood of constitutional democracy. Measures aimed at curbing FIMI and synthetic electoral fraud must remain strictly rights-respecting, transparent, proportionate, and subject to legal accountability.

Targeted counter-disinformation measures should include, among others, fact-checking and technical verification, rapid institutional clarification and algorithmic manipulation containment. Such measures should not affect democratically guaranteed rights such as press freedom, legitimate political dissent and discourse, as well as digital speech rights.

The goal is not to police opinion, regulate political debate, or restrict citizens’ freedom of expression. The sole objective is to protect the authenticity of operational election information so that when a citizen makes a choice at the ballot box, that choice is based on genuine facts rather than engineered deception.

8. Operationalising the Action Points from this Summit

This national summit comes at an opportune moment in our electoral calendar. It must not become just another conference declaration that gathers dust on library shelves. It must serve as a functional operational agreement among key stakeholders.

Specifically, INEC looks forward to the summit delivering three practical outcomes:

Clear Inter-Agency Escalation Protocols: Establishing unambiguous co-created operational guidelines detailing who detects, who verifies, who escalates, and who communicates when a major digital information crisis occurs during critical election windows. It’s vital to ensure unrelenting collaboration among all key stakeholders in this regard.

Institutionalised Fact-Checking Desks: Formalising direct, real-time communications channels between the Commission’s information officers, independent media monitoring groups, and certified fact-checking organisations.

Simulation and Crisis Stress-Testing: Moving beyond conceptual discussions by subjecting our technical and communications teams to practical tabletop exercises and simulations of AI-driven electoral scenarios ahead of 2027. The Commission, in collaboration with one of our development partners, had recently concluded a major workshop on Strategic Communication where some of these AI-driven election-day information manipulation scenarios were simulated and addressed.

Conclusion

Building credible elections requires continuous adaptation. Just as we upgraded our physical processes- from manual registers to biometric voter identification and electronic result viewing portals- so must we strengthen our digital defences to protect the information space surrounding our elections.

The integrity of the ballot in 2027 depends not only on the physical security of the polling unit, but on the authenticity of the information environment in which voters make their decisions.

As we open this summit, let our goal be certain: to ensure that by 2027, the Nigerian voter is empowered by information, not paralysed by it.

On behalf of the Independent National Electoral Commission, I extend our gratitude to the European Union, IPC, CEMESO, and DAI for organising this timely summit. I look forward to the actionable strategies that will emerge from this gathering.

Thank you, and I wish us all fruitful deliberations.

Professor Joash O. Amupitan, SAN

Chairman, Independent National Electoral Commission (INEC)

Abuja, Nigeria

Monday, 2nd November 2026

Please follow and like us: