Keypoints:
- AI is reshaping cancer detection and treatment
- Africa needs locally validated cancer AI tools
- Strong regulation must match rapid innovation
CANCER has long been one of humanity’s most formidable health challenges, leaving its mark on families, communities and entire health systems. Its burden is global, but its consequences are particularly severe where diagnosis comes late, specialist care is scarce and treatment remains beyond the reach of many patients.
Yet the fight against cancer is entering a profoundly different era.
Artificial intelligence, once discussed largely as a futuristic technology, is increasingly becoming part of cancer research and clinical medicine. From interpreting medical images and analysing tumour genetics to supporting drug discovery and treatment planning, AI is helping researchers and clinicians process information at a scale and speed that would previously have been impossible.
This does not mean algorithms are about to replace oncologists, radiologists or pathologists. Nor does it mean technology has suddenly solved cancer. What it does mean is that medicine now possesses a powerful new set of tools that could improve decisions, reduce delays and extend specialist capabilities — provided they are used responsibly.
Earlier detection could change outcomes
Early diagnosis remains one of the strongest weapons against many forms of cancer.
WHO says approximately 38 per cent of cancers can currently be prevented through existing evidence-based measures, while many cancers have a high chance of cure when detected early and treated appropriately. That distinction matters: prevention reduces the likelihood of cancer developing, while early detection seeks to identify disease at a stage when treatment is more likely to succeed.
AI is potentially strengthening that second line of defence.
One widely cited study published in Nature in 2020 evaluated an AI system for breast cancer screening using mammograms from Britain and the United States. Researchers reported reductions in both false positives and false negatives compared with conventional interpretation, although the scale of improvement differed between the two datasets.
The significance is not that machines have universally become ‘better than doctors’. Rather, AI can provide another analytical layer, helping specialists identify patterns that may otherwise be missed and potentially allowing clinicians to concentrate more attention on difficult cases.
The US National Cancer Institute has similarly stressed that AI appears particularly valuable as an adjunct to medical professionals rather than as a standalone replacement.
Pathology is seeing comparable advances. Digitised tissue slides can be examined using machine-learning systems capable of detecting patterns across enormous numbers of cells. AI is also being explored for tumour classification, risk assessment and the identification of biomarkers that could help guide treatment.
The opportunity is considerable, but clinical validation must remain the standard.
Precision oncology gets more precise
Cancer is not a single disease. Even tumours arising in the same organ can behave very differently depending on their genetics, molecular characteristics and surrounding biological environment.
That complexity is precisely where AI becomes powerful.
Modern oncology generates extraordinary quantities of information: genomic data, imaging, pathology, laboratory results and clinical histories. No individual clinician can manually interrogate every possible relationship within those datasets.
Machine-learning systems can help identify patterns linking tumour characteristics to prognosis or potential therapeutic response. In precision oncology, such technologies may support clinicians as they match particular patients with targeted therapies based on biomarkers and other molecular characteristics.
AI is also being investigated in radiotherapy planning, where computational systems can assist clinicians in identifying tumour boundaries and protecting healthy tissue. Similar techniques are being explored to predict treatment toxicity and improve decisions around systemic therapies.
Immunotherapy offers another important frontier. Researchers are using computational approaches to study tumour microenvironments and biomarkers that may help predict which patients are more likely to benefit from particular treatments.
These developments should not be confused with autonomous medicine. The strongest model remains human expertise augmented by computational capability.
AI can process enormous quantities of information. The clinician must still interpret that information in the context of the individual patient.
Drug discovery enters a new phase
The traditional development of a new medicine can take many years and requires enormous financial investment, with many potential treatments failing before reaching approval.
AI could make parts of that process more efficient.
Machine-learning and generative AI systems are increasingly being used to identify biological targets, screen chemical compounds and propose new molecular structures. Advances in protein-structure prediction, including technologies such as AlphaFold, have also expanded researchers’ ability to understand proteins associated with disease.
But claims that AI can simply produce a cancer drug within weeks and bypass conventional development should be treated cautiously.
Every potential therapy still has to satisfy demanding standards of laboratory investigation, preclinical research and human clinical trials before it can become an approved treatment.
The more realistic promise is acceleration rather than elimination of the scientific process.
AI can help researchers decide where to look, identify promising candidates earlier and potentially reduce wasted effort. In cancer research, where patients urgently need new treatments, even modest reductions in development time could matter enormously.
Africa cannot remain a spectator
For Africa and the wider Global South, the debate carries particular urgency.
The International Agency for Research on Cancer estimated more than 1.18 million new cancer cases and about 764,000 cancer deaths across Africa in 2022. Global projections also indicate that countries with fewer healthcare resources could face some of the steepest proportional increases in cancer incidence over coming decades.
That makes technological inequality a public health issue.
AI could help countries address shortages of specialist radiologists, pathologists and oncologists, particularly when incorporated into carefully designed telemedicine and diagnostic systems.
Research into smartphone-assisted cervical screening, for example, demonstrates the potential of digital imaging and machine learning to support screening in settings where conventional laboratory and specialist infrastructure is limited. Pilot studies in India and other low-resource environments have shown encouraging results, although larger clinical validation remains necessary.
For African countries, the real prize is not merely importing algorithms developed elsewhere.
Africa must participate in creating them.
Systems trained predominantly on European, North American or Asian populations may not perform equally well when applied to African patients. Different populations, disease patterns, imaging environments and healthcare systems can affect performance.
African cancer registries therefore need investment. Universities need computing capacity. Hospitals need secure digital infrastructure. Researchers require access to representative datasets, and African governments must establish rules governing how those data are collected and used.
Without this foundation, AI could widen rather than close existing health inequalities.
Responsible AI must come first
The transformation of cancer care cannot be measured only by what an algorithm can technically achieve.
We must also ask whether patients can trust it.
Cancer datasets may contain demographic imbalances. Poorly representative training data can produce systems that perform better for some groups than others. Privacy is another critical issue because genomic and health information is among the most sensitive data any individual can possess.
Regulators are beginning to confront these challenges.
Britain’s Medicines and Healthcare products Regulatory Agency has been testing approaches to AI medical-device regulation through its AI Airlock programme. Its second phase concluded in 2026, with lessons from the regulatory sandbox helping to inform the emerging framework for AI in healthcare.
The European Union is also bringing medical AI within the wider architecture of its AI Act, although important obligations for high-risk systems, including certain medical-device applications, are being introduced according to a phased timetable.
Regulation should not be regarded as an obstacle to innovation. In medicine, credible regulation is what allows innovation to earn public confidence.
Patients need evidence, not hype
Individuals also need a realistic understanding of what this technological revolution means.
People should continue to follow evidence-based screening recommendations appropriate to their age, risk profile and national health guidance. Anyone experiencing concerning symptoms should seek qualified medical assessment rather than relying on an AI chatbot or consumer application for diagnosis.
Where AI-supported clinical tools are available, patients should be entitled to ask how those technologies are validated, what role they play in decision-making and how their health data are protected.
Health systems, meanwhile, should resist the temptation to purchase technologies simply because they carry an AI label.
The question should always be simpler: does this technology improve care?
The human-AI partnership
Artificial intelligence is no longer a distant possibility in oncology. It is already influencing cancer research, medical imaging, pathology, precision medicine and the search for new treatments.
But cancer will not be defeated by computing power alone.
The future depends on combining machine intelligence with human judgement, high-quality science, functioning health systems and equitable access to treatment.
For Africa and the wider Global South, this moment presents both an opportunity and a warning. AI could help countries leapfrog some traditional healthcare constraints, extend specialist knowledge and accelerate diagnosis. But without investment in infrastructure, local research, representative data and governance, the same technologies could deepen existing divisions.
The goal should therefore not be AI for its own sake.
It should be better cancer care.
Used intelligently, ethically and inclusively, AI can help clinicians detect disease earlier, understand it more deeply and choose treatments more precisely. That may not silence cancer entirely, but it can give humanity a stronger voice in confronting it.
Professor Ojo Emmanuel Ademola is a Professor of Cybersecurity and Information Technology Management, global education advocate, Chartered Manager, UK digital journalist and Contributing Editor to Africa Briefing, strategic adviser, public intellectual, African governance thinker and General Evangelist of Christ Apostolic Church Nigeria and Overseas
