Expert Sees Potential for AI in Pandemic Response, but Warns of Ethical Concerns
Experts Discuss Key Takeaways and Challenges from Inspire Peach Project Workshop on AI in Covid 19 Surveillance
Lilongwe, Malawi - In an exclusive interview with AfricaBrief editor-in-chief, Winston Mwale, Dr Sylvia Muyingo, an Associate Research Scientist at APHRC, shared insights from a recent Inspire Peach Project workshop on the role of artificial intelligence (AI) in pandemic surveillance.
The workshop, which was organized by APHRC from June 26th and 27th at the Mount Soche Sunbird Hotel in Blantyre, brought together representatives from ministries of health, digital health organizations, civil society, and academia to discuss the use of artificial intelligence (AI) for disease surveillance.
Dr Muyingo was the Principal Investigator during the project.
Winston Mwale: What were the key takeaways from the recent event held in Blantyre?
Dr Sylvia Muyingo: There were a number of key takeaways from the workshop. First, we learned that there is a need for better coordination and governance of disease surveillance systems. This is important in order to ensure that data is shared effectively and that privacy and confidentiality concerns are addressed.
Second, we learned that there is a need to improve the quality of data used for disease surveillance. This means ensuring that the data is accurate, timely, and complete.
Third, we learned that AI can be a valuable tool for disease surveillance. AI can be used to analyse large datasets to identify patterns and trends that would not be visible to the naked eye. This information can then be used to track the spread of disease, predict outbreaks, and target interventions more effectively.
Winston Mwale: How did the participants find the event?
Dr Sylvia Muyingo: The event was well-received by the participants, who showed great interest in involving stakeholders from different sectors and improving health systems' data sources. There were discussions on building a working group to continue the conversations and capitalise on existing synergies.
Winston Mwale: What challenges and opportunities did you face while organising the event?
Dr Sylvia Muyingo: One of the challenges we faced was in reaching out to stakeholders and getting them to attend the workshop. This was due to a number of factors, including travel restrictions and the need to obtain visas.
Another challenge we faced was ensuring that the workshop was accessible to all stakeholders. This meant providing interpretation services in multiple languages and making sure that the workshop materials were accessible to people with disabilities.
Despite these challenges, we were able to organise a successful workshop that brought together a diverse group of stakeholders. This was an important opportunity to share ideas and discuss how we can use AI to improve disease surveillance in Malawi and beyond.
Winston Mwale: What lessons can be learned from this event and applied to future projects?
Dr Sylvia Muyingo: One of the key lessons we learned is that it is important to build trust with stakeholders. This means being transparent about the purpose of the project and ensuring that the data is used in a responsible way.
Another lesson we learned is that it is important to be flexible and adaptable. Things don't always go according to plan, so it is important to be able to adjust as needed.
Finally, we learned that it is important to have a clear vision for the project. This will help keep everyone focused on the goal and make it more likely that the project will be successful.
Winston Mwale: How do you see the potential of AI and data science tools in pandemic response or other situations?
Dr Sylvia Muyingo: AI has great potential in pandemic response, especially in collecting and analysing data for real-time decision-making. It can improve efficiency in generating knowledge and reports, facilitate medication delivery, and assist in diagnosis. However, AI development in the African context requires more data specific to the region.
Winston Mwale: The whole project you worked on was based on AI. How do you envision the potential of AI and data science tools in responding to crises like COVID-19?
Dr Sylvia Muyingo: AI holds tremendous potential, particularly in pandemic responses where vast amounts of data are involved. It can be used not only to collect data but also to analyse it, thereby informing and improving decision-making in real-time.
By leveraging the lessons we've learned from previous experiences and the massive influx of data during a pandemic, we can enhance the efficiency of generating knowledge from the collected data.
This, in turn, enables us to generate reports that aid decision-making and better prepare us for future outbreaks. Additionally, AI can be instrumental in delivering medications and expediting diagnoses, relieving some of the burden faced by healthcare systems during crises.
In the African context, AI can play a crucial role, but we need more data from our specific context to improve the adaptability of AI tools.
Winston Mwale: The quality and quantity of data are central to AI's effectiveness. What ethical considerations should be in place when using AI in healthcare?
Dr Sylvia Muyingo: Ethical concerns arise when we consider that AI relies on the data we provide. For instance, in our project, we found that the available data mostly came from hospitals. However, many people in our context cannot afford to visit hospitals or simply do not have access to healthcare facilities.
Consequently, these individuals are excluded from the data, introducing bias.
Additionally, if we use social media as a data source, we must acknowledge that younger individuals are more likely to engage on these platforms, potentially skewing the data further. Inequalities within the data source will be reflected and amplified by AI.
This highlights the need for ethical considerations regarding privacy, inclusion, and the reinforcement of existing biases. Transparency and accountability are crucial in AI development to ensure public awareness and participation.
By fostering a safe and secure environment and clarifying the purpose of data usage, we can build trust among citizens.
Key takeaways from the event:
Participation from various stakeholders, including Ministries of Health, digital health organisations, civil society, and universities.
Focus on data-related issues, improving data quality for surveillance, and the need for data-sharing policies.
Integration of population health data into the surveillance system and the importance of community surveillance.
Capacity building for technical work and making data actionable for policy decisions.
Coordination and governance improvements, including multidisciplinary collaboration and the incorporation of the One Health approach.
Investment in infrastructure and the recognition that AI is here to stay.
Participant reactions to the event:
Participants found the event well-received and expressed interest in involving stakeholders from different sectors.
Emphasis on supporting health systems, improving data sources, and building capacity at the primary data collection level.
Suggestions for building a working group to continue discussions and synergies.
Challenges faced during event organisation:
High-level delegates' commitments and government duties affected attendance and required delegation.
Time-consuming processes for reaching out to stakeholders and arranging travel logistics and paperwork.
Technical issues during presentations that required troubleshooting and some delays.
Lessons learned from the event and study:
The need for a federated approach to data sharing, where data stays with data providers, and services/tools are shared.
Understanding that sharing COVID-19 data involves addressing ethical, privacy, and confidentiality concerns.
Data does not need to be standardised but should conform to a minimum data specification.
Emphasis on communication, collaboration, and reaching agreements with multiple stakeholders.
Acknowledgement that finding solutions may require trying different approaches and exploring various data sets.
The Potential of AI and Data Science Tools in COVID-19 Response:
AI has great potential in pandemic response due to its ability to collect and analyse large-scale data.
AI can improve decision-making in real-time and enhance efficiency in generating knowledge from collected data.
AI can be used for delivering medications, diagnosing, and improving healthcare systems.
African context-specific data is essential for developing and adapting AI tools.
Ethical considerations for AI use in healthcare:
AI reflects the biases and inequalities present in the data it is based on.
Data sources and technology used may result in data biases and exclusions.
Ethical issues arise when privacy is compromised or when inequalities are reinforced by AI.