Hospitals and Chatbots: How AI is rewriting medical consultation in Nigeria

GIFT HABIB examines how Nigeria’s strained healthcare system is quietly shifting first medical consultations from clinics to Chatbots, as patients increasingly turn to Artificial Intelligence for answers.

In the first weeks after she had her baby, 26-year-old first-time mum Victoria was still learning how to read the smallest changes.

Some days, it was the crying; the long stretches of it, especially at night. Other times, it was something as simple as stool colour that left her uneasy. Everything felt new, and with that came a steady stream of questions she did not always have answers to.

In those moments, access to medical help was not always immediate. The nearest Primary Healthcare Centre in Oshodi, Lagos State was quite a distance away, and there were also concerns about the level of care available. One night, when the situation became particularly serious, she reached for her phone, opened the ChatGPT app, and began typing her question.

“Honestly, that night, October 3rd, 2025, was one night I will not forget. I was really tired and stressed. So I reached out to my phone as per an educated and digital woman,” she laughed as she narrated.

Her questions were simple but urgent.

“My newborn has green stool, is it normal? “Why is my baby crying and not sleeping at night?” “What can I do if my baby has gas?” she revealed what she asked ChatGPT.

Then Victoria continued, “ChatGPT helped me figure it out. It gave me responses on what caused it, what I can do, and how to manage the situation.

“I did not use AI when I was pregnant or dealing with fibroids. But postpartum, it became my guide.”

There were also moments when the stakes felt higher. Once, she bought malaria medicine for her baby but realised she no longer had the prescription details. Rather than go back, she turned again to her phone, entered her child’s weight, and asked for dosage guidance.

She added, “A few days later, I spoke with a paediatrician and described the prescription to her, and she confirmed it was correct.”

It was not just about convenience. In that moment, it was about making a decision without delay and hoping it was the right one.

A few of her questions had nothing to do with the baby. They were about her own body.

She added, “Also, after giving birth, my bleeding lingered for more than three weeks. I thought something was wrong. ChatGPT helped me understand that it was normal, even up to six weeks. ChatGPT really helped me through that.”

Her baby has grown. For Victoria, the appeal was simple. Answers came instantly, even when professional help was not immediately available. The trade-off, though unspoken at the time, was that speed replaced verification.

 Before the doctor, there is the phone

In Nigeria, the growing use of AI chatbots is changing how people seek information, including on health-related concerns. What was once a last resort; searching symptoms online has become part of early decision-making.

A 2026 Google–Ipsos survey found that 88 per cent of adults in Nigeria surveyed have used AI chatbots, higher than the global average of 62 per cent. The findings indicate a rapid and widespread adoption of AI tools in Nigeria, extending beyond access to everyday use across various aspects of life.

In a healthcare system characterised by long waiting times, uneven access, and high out-of-pocket costs, AI offers a more immediate solution. It provides answers without queues, explanations without consultation fees, and a level of privacy many users value. That speed is part of its appeal. It is also where the tension begins.

Trusted with caution

For Wisdom, the 34-year-old banker said when something feels off in his body, he does not rush straight to a hospital. He checks his phone first.

“I check AI summaries when I am worried about symptoms. Google now has AI assistants that summarise information, leaving the possible causes and remedies. Before, I would have to sift through countless links,” he said.

But Wisdom is careful not to treat it as final.

“But I always remember the AI disclaimer; it is not a medical expert. So if it looks serious, I confirm with a pharmacist or doctor. Sometimes, I follow the suggestions if they make sense, but I never rely solely on AI for treatment,” he adds.

On the other hand, a university student, Favour said it started with something she was not even trying to solve on her own. A laboratory test had already raised concern, and she said she needed something that helped her understand what it meant.

The student explained, “AI has helped me understand my body without seeing a doctor directly. When a laboratory result revealed that I was pre-diabetic, ChatGPT confirmed it and suggested dietary changes and supplements.”

She said it also became useful for something she had lived with for years without real improvement.

“For dandruff, which I have had for almost ten years, AI analysed the ingredients in a cream my vendor wanted me to use and advised against it. It prescribed a better alternative. It worked immediately,” she explained.

But even with those experiences, she draws a clear line. “Of course, AI cannot replace a doctor; especially for serious conditions, like pregnancy or emergencies,” the student cautioned.

For Chioma, her own use comes from something simpler than curiosity or experimentation. It is access or the lack of it.

“I use AI for health-related issues because of the queue in our public health systems.

“The General Hospital around my area is nothing to write home about,” she explained.

She said there were moments when it was not even about treatment, just mental preparation before stepping into a hospital.

Chioma admitted, “I did not want to hear bad news or to be diagnosed. So I use AI to prepare my mind ahead of my hospital visit if it is severe. But a few times, what I see in ChatGPT can give me high blood pressure especially the symptoms.”

Christopher is not any different. It is something he has repeated enough times for it to become routine.

“I do use AI to check for diagnosis. And most of the time, if not all the time, when I feel certain symptoms, I just type out the symptoms, how I am feeling and I get the responses.”

He said what comes back is usually enough to decide the next step. “I go to the hospital afterwards. I tell them my symptoms and the lab personnel would run me a test and nine out of 10 times, it came out exactly as what AI told me.”

Looking across these accounts, a pattern starts to show. AI isn’t replacing hospitals, but it is quietly changing how and when people decide to seek care.

More often, people are making those decisions sooner, quicker, and with a greater sense of independence than before.

Helpful or harmful?

Globally, AI tools are now used by hundreds of millions of people. Major platforms such as OpenAI report hundreds of millions of weekly users for ChatGPT alone, reflecting the rapid adoption of generative AI. Industry research, including reports from McKinsey, similarly shows that AI adoption has expanded quickly across organisations worldwide, pointing to its growing integration into work processes and everyday information-seeking behaviours.

As a result, users now engage AI chatbots for a wide range of tasks, including information seeking, writing assistance, and practical guidance, alongside other everyday uses. This shift has been gradual but transformative. According to an article published in The Conversation, “When ChatGPT became publicly available on Nov. 30, 2022, the shift felt sudden, but it was built on years of incremental integration.

“AI’s presence is now so mundane that people consult chatbots for recipes, use them as study partners and rely on them for administrative chores.”

Within this expanding use, health-related queries have also emerged as part of how people interact with AI systems. Analysts broadly describe this evolution as a transition from static search engines to conversational AI systems, where users increasingly rely on tools that interpret, summarise, and generate responses in real time.

However, concerns about reliability particularly in sensitive areas like healthcare are growing. Studies have shown cases where chatbots provided misleading or overly confident medical advice. For example, a Nature Medicine evaluation reported in 2026 found that some AI systems failed to recognise urgent medical emergencies and often advised users to delay seeking care, raising concerns among healthcare professionals about overreliance on these tools.

These concerns are also reflected in Nigeria. A growing number of studies and professional assessments indicate that while artificial intelligence tools are being explored in healthcare, questions remain about safety, accuracy, and misuse. Local research among medical professionals shows that although many doctors and health workers are open to AI-assisted tools, they remain cautious due to risks such as misinterpretation of patient symptoms, lack of regulatory oversight, and overreliance on systems not fully adapted to local health realities.

In 2024, public health experts under the Health Promotion and Education Alumni Association of the Ibadan College of Medicine warned Nigerians against using artificial intelligence for self-diagnosis or medical prescriptions, saying it could be harmful to health. They expressed concern over the growing trend of people turning to the internet instead of qualified healthcare professionals, noting that AI-generated or online medical advice may fail to consider a patient’s medical history and could therefore do more harm than good.

This caution was echoed during the same engagement. During the conference, the President of the Ibadan College of Medicine Alumni Association Worldwide, Professor Emmanuel Otolorin, during the two-day Annual General Meeting and Scientific Conference of the HPEAAICM, said that although AI presents many opportunities that can be harnessed, public health consumers and practitioners must be aware that AI can hallucinate and create what doesn’t exist.

“My son is a computer scientist, and he actually showed me how an AI tool created a drug that doesn’t exist and also described side effects of a drug that actually doesn’t exist. So AI can hallucinate. You have to be a content expert to know what it’s talking about.

“If you use an AI tool and ask it to do something, you have to be knowledgeable enough to know that what it has done is correct and not just adopt it.”

Strained healthcare infrastructure

Nigeria’s healthcare system continues to face structural pressures that affect how quickly and easily people can get care. According to the Nigerian Medical Association and widely cited workforce estimates, there are roughly 30,000 to 55,000 practising doctors serving a population of over 200 million. This deficiency shows up in everyday life through long waiting times, uneven access to doctors, and limited availability of timely consultations, especially outside major cities.

Primary healthcare centres, which are supposed to be the first point of contact for most patients, are often stretched thin. Many deal with shortages of basic medicines, diagnostic tools, and staff, while others struggle with inconsistent funding and infrastructure issues like unreliable power and limited equipment. Because of this, patients often skip primary care entirely and head straight to bigger hospitals, which are already overcrowded.

These pressures are made worse by cost. Out-of-pocket spending still accounts for a large share of healthcare payments in Nigeria. For many households, a hospital visit means more than just consultation fees, it can include transport, tests, and medication. So when symptoms do not feel urgent or clear, people often delay going in or look for other ways to understand what’s happening first.

In that gap, digital tools like AI chatbots and search-based health assistants are increasingly being used as a first point of reference. They do not replace doctors, but they do offer quick explanations and suggestions that help people make sense of symptoms while they figure out their next step.

The growing reliance on artificial intelligence for everyday health decisions is no longer happening only at an individual level. It is beginning to take shape within public policy. In December 2025, the Speaker of the House of Representatives, Abbas Tajudeen, announced plans for the National Assembly to fund AI-powered healthcare subscriptions for persons with disabilities over the next five years, through the Koyo Navigate platform. The initiative was designed to expand access to doctor-linked consultations, AI-guided health information, and other digital health services tailored to people with disabilities.

It reflected a shift already underway across the country, where Nigerians turn to AI tools as a first step in understanding symptoms before seeking professional care. By backing a structured, supervised platform, policymakers appear to be responding to this behavioural change, attempting to channel informal AI use into a more regulated and clinically supported system. For persons with disabilities, who often face additional barriers to accessing healthcare, such tools could provide quicker and more consistent access to medical guidance.

However, the integration of AI into formal healthcare delivery also raises important questions. While platforms like Koyo combine artificial intelligence with input from medical professionals, they still rely on digital infrastructure, user interpretation, and data that may not always fully reflect local health realities. Concerns around accuracy, oversight, and data privacy remain, particularly as AI systems become more embedded in decision-making processes that affect patient outcomes.

Government action

Last week, the Federal Government pledged to accelerate the adoption of digital technologies across Nigeria’s health sector as artificial intelligence tools and other digital platforms become more widely used by citizens for first-line medical advice.

The Minister of State for Health and Social Welfare, Adekunle Salako, who spoke at a World Health Day event in Abuja, said the government is prioritising innovations such as telemedicine, interoperable electronic health records, and a national health information exchange system to improve access to healthcare services.

Salako noted that Nigeria is determined to keep pace with global health innovation, noting that technology is transforming healthcare globally from telemedicine to artificial intelligence in diagnostics and digital health records.

He acknowledged that the sector continues to face structural challenges, including workforce shortages, weak infrastructure, and gaps in local production of medicines and vaccines, stressing that reforms must be guided by science and evidence.

Salako also stressed equity in healthcare access. “We are clear in our vision for the country’s health sector, where every Nigerian, regardless of income level, geographic location, or social status, should have access to quality healthcare services when they need them, without suffering financial hardship,” he said.

He added that ongoing reforms are focused on strengthening primary healthcare, improving supply chains, and expanding access to essential services, particularly in rural communities, with initiatives such as the Power for Health programme aimed at improving electricity supply in health facilities.

Learning how to respond

Health practitioners have repeatedly warned against self-diagnosis using AI systems, noting that such tools are not designed to replace clinical evaluation.

A public health practitioner at the Public Health Department, Federal Capital Territory Health Services and Environment Secretariat, Sunday Awobanjo, said “concerns around misinformation are making some health workers cautious about AI-generated medical advice”.

Awobanjo noted that while scepticism exists, AI could still play a supportive role in healthcare delivery.

According to the microbiologist, AI may assist clinicians in diagnosis by helping them “go deeper” and consider possibilities that are not immediately obvious. Awobanjo added that it could also support medical decision-making in emerging diseases that are not always fully covered in formal training.

He further suggested that AI could be useful in public health practice, particularly in areas such as disease surveillance, climate-related health risks, and access to global health information.

However, he expressed uncertainty about how far “government investment and structured policy have gone in preparing health workers for AI integration.”

A medical practitioner at Wuse District Hospital, Abuja, and the President, Association of Resident Doctors, FCT Chapter, Dr George Ebong, explained that AI tools like ChatGPT can help patients understand general health information but are not reliable for “diagnosis or treatment decisions.”

Ebong explained that while AI can generate useful explanations, “it does not examine patients, conduct tests, or fully account for individual medical history.” As a result, it can sometimes miss serious conditions or provide incorrect conclusions, he added.

Ebong also pointed out a structural limitation. “Many AI systems are trained on data outside Africa, meaning they may not accurately reflect common local conditions such as malaria or typhoid, or the realities of Nigeria’s healthcare system.”

Still, he acknowledged their value as a first layer of information, particularly in helping patients understand symptoms and prepare for medical consultations.

“AI should be used to inform patients, not to replace doctors.”