INEC ICT director struggles to explain how AI-generated photos entered voters’ register
By Ishaya Ibrahim
The Independent National Electoral Commission (INEC)’s Director of Information and Communication Technology, Lawrence Bayode, struggled to explain how photographs generated using Artificial Intelligence (AI) tools found their way into the commission’s recently published official voters’ register.
Bayode, who appeared as a guest on Arise News’ The Morning Show, was responding to questions about how photographs confirmed to have been generated using AI tools ended up in the register.
Edo State, where Governor Monday Okpebholo assured President Bola Tinubu of 2.5 million votes in the 2027 election, is the state most affected by the AI-generated photographs in the voters’ register.
When asked by the programme’s interviewer, Rufai Oseni, how the AI-generated photographs made it into the register, Bayode struggled to provide a clear explanation.
The exchange went as follows:
Rufai: How did the AI pictures get into the register? Secondly, can you show me a document outlining INEC’s AI usage policy that has been shared widely with the public?
Bayode: When you say AI use… Let me answer the first question first while I try to pull up the policy guiding us, even though we have not deployed AI in our voter registration process. But I can tell you how the AI-generated photos found their way into the register.
First, you know we have registration officers in the field, and these registration officers have login details to access the system. What we have found is that some of these registration officers, not only in the places we are seeing on social media, have engaged in different activities that introduced these irregularities into the register.
Some of them, not all, across the country, have done different things to bring these irregularities onto the register.
Like I said, this happened across the country. We were dealing with this even before we released the preliminary register. The AVs detected a lot of these irregularities.
We were working against time. You know the time was short, and we were working hard to ensure we met the 90-day deadline for releasing the preliminary register.
So, we were cleaning up the register and identifying a lot of irregularities. Different people were doing different things, and some of the irregularities were similar.
Yes, you are very correct. Some of the pictures were AI-generated, while others were not. Some were simply the result of very poor registration practices.
These images got into the register through the registration officers because they have access to register people.
When the AVs were run on the system, we identified these people. We know what they did and how they did it, but, honestly, I don’t know why.
