Nigeria Needs Trusted AI in Healthcare — Experts

The Africa Hub for Innovation and Development (AFHID) has called for the development and adoption of trusted, homegrown and human-centred Artificial Intelligence (AI) solutions to tackle healthcare challenges in Nigeria.

The call was made on Thursday at a pre-summit webinar on AI in health, where experts examined the opportunities and challenges of deploying emerging technologies across different areas of the health sector.

The Chief Executive Officer of Advantage Health Africa (Aha), Ms Abimbola Adebakin, said trust was especially critical when deploying AI in healthcare, describing it as a "low-forgiveness sector." She explained that Aha had built its 'myMedicines' platform around familiar channels such as WhatsApp, allowing users to interact through a medium they already understood before AI was introduced into the process.

"Trust comes when people know to an extent who is behind the device," she said, adding that Aha had similarly automated order processing and customer engagement on its website because it was a channel already familiar to its users.

According to her, trust can be lost through a single confusing AI response, making it essential for innovators to build systems gradually while ensuring their operations remain understandable and reversible. She said AI was increasingly being applied beyond last-mile healthcare delivery, helping patients understand drug-drug interactions, possible side effects and proper medicine storage, while also supporting backend pharmacy operations such as demand forecasting and inventory control through the Advantage Platform.

Speaking on misinformation and deepfakes, Dr Chinonso Egemba, Co-founder of AwaDoc, said malicious actors were increasingly using AI to clone trusted faces and spread false health information. He noted that while early deepfakes were relatively easy to detect, advances in AI meant future manipulated content could become far harder to distinguish from genuine material.

Egemba, a public health communication strategist, called for greater public exposure to verified health information to help people recognise authentic content. "You cannot place the responsibility of identifying misinformation on the populace. The only way you can identify fake is if you have been exposed to the original," he said.

He cited a study by KPMG and the University of Melbourne which found that 59 per cent of Nigerians surveyed listed loss of human interaction as their top AI-related concern, stressing the need to ensure AI's expansion does not eliminate human involvement in healthcare. He argued that scale and trust should complement each other, with AI serving as a primary source of health information for large populations while human experts step in for specialised support when needed.

"What we need is human-in-the-loop design. At the point that matters, there should be help that can come in by a human being," he said, also urging health institutions to leverage trusted community structures and local languages to amplify verified health information and counter misinformation.

The Managing Director and Co-Founder of Fertitude, Dr Kieva Chris-Amusan, said women would be more likely to adopt AI-powered health solutions when the technologies were safe, private, accurate and capable of solving genuine problems. She said Fertitude keeps users' data anonymous, storing records under identification numbers rather than names or email addresses, with a privacy policy that clearly explains how user information is handled.

"The broad answer is if it works and it protects my privacy, gives me the answer I need, I will come back," she said.

Chris-Amusan noted that many large language models are trained predominantly on Western populations and may therefore fail to adequately reflect the experiences of African women, stressing the need for AI health solutions built for Nigeria to incorporate local accents and languages—including Hausa, Yoruba and Igbo—along with images and contexts familiar to Nigerian users. She added that appropriate user experience, accurate retention of patient history and product-led reminders were also important for sustaining engagement with AI health platforms.

The panellists identified poor design thinking, inadequate institutional validation and fragmented identifiers as key barriers to scaling AI solutions in Nigeria's health sector, calling for clearer data protection rules and stronger institutional backing for innovators, while encouraging stakeholders to share failures and lessons from AI projects to strengthen the wider health technology ecosystem.

They also stressed the importance of developing a single identifier for health services, citing the Bank Verification Number (BVN) system in the financial technology sector as an example of how a common identifier could support service delivery. Additionally, they called for government-led infrastructure and support mechanisms—including a health innovation accelerator modelled on approaches such as the United Kingdom's National Health Service (NHS)—to support health technology founders, noting that government is often more open to innovation than many founders perceive, provided innovators can demonstrate their solutions' value and understand the different layers of governance, infrastructure and literacy at the facility level.

The pre-summit webinar is part of activities leading up to the main AI in Health Summit convened by AFHID, which is expected to further examine the role of AI and innovation in addressing healthcare challenges and strengthening health outcomes in Nigeria.