LIRNEasia hosts Exclusive Forum in Colombo ‘Unlocking Data for AI in Sri Lanka’

Panel of Experts

For Sri Lanka, AI offers a unique opportunity to overcome traditional development hurdles, particularly in sectors like health, education, agriculture, and climate resilience

Despite AI holding great potential for addressing global challenges, developing nations face significant barriers to realising its benefits. In particular, data for AI applications are often scarce, costly, outdated, and not representative of marginalised groups. The lack of clear policies, incentives and coordinated efforts also impedes responsible collection and use of data. For Sri Lanka, AI offers a unique opportunity to overcome traditional development hurdles, particularly in sectors like health, education, agriculture, and climate resilience. The country’s new draft strategy, ‘AI Sri Lanka 2028: Sri Lanka’s National Strategy on AI’ highlights data as the foundation for AI-driven solutions and outlines initiatives to improve data availability and quality. LIRNEasia states “As the country prepares to embark on a journey of digitalisation, a discussion about using data for responsible innovation is timely”. 

LIRNEasia hosted an exclusive forum on the transformative role of Artificial Intelligence (AI) titled ‘Unlocking Data for Artificial Intelligence in Sri Lanka’ on December 7. The event highlighted various examples of both successful and unsuccessful data projects from around the globe, providing insights into the practical challenges, and the need to address ethical concerns and trade-offs in data. The session was conducted by a group of AI experts, data scientists, policymakers, and industry leaders to explore pragmatic ways to leverage data for informed decision-making and beneficial societal and organisational impact.

The Chair of LIRNEasia, Prof. Rohan Samarajiva, opened the session by highlighting the organisation’s ongoing work with AI and machine learning to address societal and policy issues. He said “We’ve been working on AI since 2012, we are a research organisation and we keep moving”.LIRNEasia conducts research in collaboration with governments and civil society on projects and the public policy developments process. LIRNEasia Team Lead for Data, Algorithms and Policy, Merl Chandana, explained that while data is crucial for AI, its practical application can be challenging. He concluded by summarising the session’s focus: “Despite these challenges, interest in AI continues to grow due to its potential. This forum will explore how to effectively obtain and use data, the risks involved, and the importance of ethical and responsible data usage.”

In his keynote speech, Dr. Adolfo De Unanue, Director of Research at the School of Government and Public Transformation and leader of the Artificial Intelligence for Public Value Initiative at Tecnológico de Monterrey, Mexico, said “Machine learning is a small yet an important part of the process to improve the lives of people”. He shared several international examples which included optimising logistics at Mexico’s Port of Veracruz and using AI to improve efficiency of criminal case resolutions. He emphasised that success requires more than data and technical skills, noting that clear goals and strong leadership are crucial. 

During the first panel discussion moderated by Merl Chandana, who raised a question regarding the key lessons learnt in trying to leverage data, Dr. Kasun Amarasinghe, senior project scientist of Data Science and Public Policy Lab, Carnegie Mellon University said “We need to broaden our focus from merely using AI to incorporating data driven and evidence driven decision making in public policy. It’s important to consider the human element because the success of these algorithmic systems depend on the action of people who use it. Thus, we need to empower the public sector with the necessary data skills and literacy to effectively use and procure these tools.” 

Furthermore, Yasith Fernando, Chief Data and Analytics Officer at Dialog Finance highlighted the importance of defining the problem statement, understanding data and its underlying processes, and the importance of an organisation’s transition to becoming data driven. “No organisation or country can be AI driven if they’re not data driven to start with” he concluded. 

As a response to the legal concerns regarding data, particularly with the newly implemented data protection Act, Ashwini Natesan, legal consultant, expert in data governance and data protection said that it is essential to distinguish between personal and non-personal data. “When using data, we must comply with the obligations set out under the Personal data protection Act such as lawful processing, purpose limitation, and time-bound data retention. When non personal data is involved, data governance principles must be considered as well.”

In summary, the first panel drew on examples from both the public and private sectors to discuss how to ensure the success of data-driven projects while safeguarding lives and adhering to regulations. Speakers emphasised the importance of capacity building, responsible data usage, and balancing data protection with public benefit.

In his speech, Prof. Udayanga Hemapala, Secretary of the Ministry of Power and Energy, discussed the transformative potential of data and AI in the energy sector. He pointed out that inadequate data storage and limited processing capabilities hinder quick response times in the sector and stressed on the need for an equitable energy transition, warning that its costs could disproportionately affect lower-income customers. Prof. Hemapala further outlined key challenges to AI-driven energy management such as data quality and availability issues, real-time decision-making, cyber security risks, and regulatory barriers. “It’s important to promote affordable clean energy, for this, more research must be conducted to identify the best suited electricity tariff and strategies to implement this form of project in Sri Lanka.”

The second panel discussion moderated by Prof. Rohan Samarajiva discussed how data and innovative approaches can transform the energy sector, particularly in managing electricity consumption and improving efficiency. In response to the hypothesis presented by Prof. Samarajiva on how low income households often use less energy-efficient appliances, leading to higher costs, Merl Chandana explained that LIRNEasia has conducted interviews with over 4,000 households to understand their energy expenses. The results revealed that lower income households spend more due to their reliance on inefficient appliances. The project combined smart meter data and surveys for a better understanding of energy usage patterns, offering valuable insights into consumer behaviour.

Prof. Moinul Zaber from the United Nations University shared how data-driven strategies in Bangladesh helped tackle electricity challenges, leading to improved grid reliability and service efficiency. Data analysis revealed that payment delinquency was not only linked to poverty but also to the distance from payment centers. Introducing online and mobile payment systems significantly reduced delinquency while a data-driven chat bot improved customer engagement by providing timely updates. He added, “With the availability of data, better decisions can be made to gain the consumer’s trust and with this, the transactional cost decreases. Open data mechanism is the ideal solution since it combines AI powered innovations as well.”

In summary, the second panel underscored the importance of cooperation between data scientists, social scientists, and policy makers to navigate the challenges of data management and policy development in the energy sector. The forum concluded with a call for collaboration among multiple stakeholders to address the challenges of data governance.