Prof Samuel Essien-Baidoo, Acting Director of Research and Development at the National Vaccine Institute (NVI), Ghana, has called for robust legal and regulatory frameworks to ensure accountability in the use of Artificial Intelligence (AI) in healthcare.
He warned that innovation without responsibility could undermine public trust and patient safety.
Speaking on the topic, “When AI Goes Wrong: Legal Perspectives on Accountability, Responsibility and Liability in Healthcare AI,” at the International Conference on Artificial Intelligence and Health (ICAIH 2026) in Ho, Prof Essien-Baidoo said AI held enormous potential to transform healthcare through faster diagnosis, precision medicine, accelerated drug discovery and improved clinical decision-making.
He, however, cautioned that the technology also posted significant risks, including misdiagnosis, biased decision-making, privacy breaches, unsafe recommendations, and undetected system failures if not properly governed.
He stressed that AI should not be regarded as an autonomous decision-maker but as a tool whose outcomes depended on the accountability of the humans and institutions deploying it.
According to him, when an AI system produced a harmful recommendation, responsibility could extend to clinicians, hospitals, software developers, regulators and governments, depending on the circumstances.
He noted that although AI could support medical practice, it lacked legal personhood and therefore could not bear legal responsibility for its actions.
Prof Essien-Baidoo explained that existing legal doctrines, including negligence, professional liability, product liability and corporate responsibility, provided a foundation for addressing AI-related harm but required adaptation to the realities of algorithm-driven healthcare.
He urged governments to adopt lifecycle regulation for AI systems, requiring transparency, validation using local data, explainability, safety testing and continuous post-market monitoring before and after deployment.
He said Ghana already had important regulatory institutions, including the Food and Drugs Authority, Data Protection Commission and professional councils, which could serve as anchors for AI governance.
He, however, advocated stronger coordination among regulators, healthcare institutions, academia and industry to establish a unified framework for oversight and accountability.
Highlighting AI’s growing role in vaccine development, disease surveillance and public health, Prof Essien-Baidoo warned that failures in AI-powered systems could affect entire populations rather than individual patients.
He therefore called for independent audits, protected reporting channels for frontline health workers and clearly defined chains of accountability involving vendors, health authorities and regulators before AI systems were deployed in national health programmes.
He proposed a five-pillar framework for responsible AI in Ghana’s health sector, centred on human accountability, transparency, safety, fairness and continuous oversight.
He further recommended the development of a national AI governance policy for health, certification pathways for clinical AI applications, AI oversight committees in hospitals, integration of AI law and ethics into higher education curricula and promotion of responsible innovation by industry.
Prof Essien-Baidoo said governance should not be viewed as an obstacle to technological advancement but as the foundation for safe and sustainable innovation.
“Artificial Intelligence may inform decisions, but accountability must always remain human,” he said, adding that public trust would remain the currency for successful adoption of AI in healthcare.
