Empowering Cambodia through Artificial Intelligence (AI) and Data: insights from the human-centered AI adoption conference at ASEAN COSTI-85
On Monday, June 3, 2024, ODC attended an afternoon conference titled “Enabling Environment for a Human-Centered Adoption of AI in ASEAN” at the Sofitel Angkor Phokeethra Resort in Siem Reap, Cambodia. The conference was a side event of the larger 85th meeting of the ASEAN Committee on Science, Technology, and Innovation (COSTI-85). The conference was sponsored by the Cambodian Ministry of Industry, Science, Technology & Innovation (MISTI) in collaboration with The Asia Foundation and Stanford Human-Centered Artificial Intelligence (HAI). The event covered a wide array of artificial intelligence (AI) topics pertinent to policy makers, private-sector stakeholders, educational institutions, and civil society organizations in ASEAN. The conference follows in the footsteps of the Cambodia Digital Government Policy 2022-2035 released by the Ministry of Post and Telecommunications (MPTC) in January 2022 and the Digital Tech Roadmap published by MISTI and the National Council on Science, Technology, & Innovation (NCSTI) in July 2023. The Digital Tech Roadmap identified AI and machine learning as the highest-scoring key technology in terms of both importance and feasibility when evaluated against other emerging technologies. The Roadmap seeks to utilize AI to develop “a vibrant ecosystem enabling digital business in the country.” Additionally, the Cambodia Digital Government Policy 2022-2035 seeks to accelerate the availability of data that could be used as AI training data through the digital governance and open data initiatives. The conference aligns seamlessly with ASEAN’s Digital Economy Framework Agreement (DEFA) by promoting ethical AI governance, fostering regional collaboration, and enhancing digital integration to drive economic growth and innovation across member states.
The first speaker from Stanford HAI presented on the topic of fairness in generative AI. The presentation highlighted several contemporary issues in AI including a lack of global equity and inclusion, bias in decision making, and algorithmic unfairness. Prior research revealed gender bias in popular large language models when used in hiring decisions. The speaker emphasized the need for socio-technical solutions that recognize the inherently subjective nature of computer systems and algorithms that are embedded with human biases. The importance of grounding AI in local context was exemplified by his team’s development of a large Vietnamese language model that is publicly available online for free.
The second speaker of the afternoon was also from Stanford HAI. The research focused on human expectations of AI systems. The presentation focused on the trustworthiness of AI systems. The speaker pointed to a clear distinction between a “pilot” and “passenger” mindset towards AI in the general public. A “pilot” is eager to adopt AI systems and feels in control when utilizing AI systems. A “passenger” is wary of AI systems because they feel that they surrender too much control to the systems. The data indicates that business managers are more likely to trust AI systems than employees which may hinder effective adoption. Lastly, the speaker re-iterated that privacy is fundamental to building trustworthy AI systems.
Afterwards a virtual panel presented on some of the most popular AI regulations and governance models around the globe. A speaker from the Université Côte d’Azur emphasized the importance of utilizing technology to jumpstart economic growth for developing countries. He also mentioned the ongoing debate on whether the benefits of AI technology will diffuse among a wide base or whether it will merely enrich the current dominant entities with large market power. The need to govern AI as a system rather than as an individual piece of technology was emphasized. This would be especially important if any significant progress was made regarding global governance of AI. The speaker proposed governing AI in four dimensions: competition policy; industrial policy; IP policy; and dual use technology.
Following these remarks, the individual panel member gave an overview of the AI regulations and governance in the European Union, the United States, and ASEAN. The European Union adopted its AI Act in May 2024. It seeks to provide standardized AI definitions, enable safe and trustworthy AI systems, promote innovation, and safeguard fundamental rights. A speaker from Center for Digital Trust (C4DT) discussed the European Union’s risk-based approach to AI safety that classifies AI systems into four risk categories: unacceptable risk; high risk; limited risk; and minimal risk. Subsequently, a researcher from Stanford HAI detailed the United States Executive Order on AI that provided guidance to federal agencies. He also summarized some of the AI bills in Congress that have been proposed. Lastly, a representative from the AI Asia Pacific Institute (AIAPI) detailed proposals from the ASEAN Digital Minister’s Guide on AI Governance and Ethics that were endorsed in February 2024. Singapore’s AI efforts were specifically praised and highlighted. The speaker noted Singapore’s efforts to develop technical standards that are interoperable with the United States’ NIST.
The country representative in Cambodia for the Asia Foundation, spoke after the conclusion of the virtual panel. She highlighted the need for fair and human-centered AI adoption that enables technology to thrive and serve broader societal goods. She encouraged collaboration and dialogue on AI in order to build instructional capacity.
Concluding remarks were given by His Excellency Mr. Hem Vanndy, Minister of Industry, Science, Technology & Innovation, who extolled the importance of continuing the conversation amongst the relevant stakeholders.
Role of AI in Cambodia’s Small and Medium Enterprises (SMEs)
The conference emphasized the collaborative efforts required between the public sector, private sector, academia, and civil society organizations in order to successfully utilize AI to expedite the development of Small and Medium Enterprises (SMEs) in Cambodia. SMEs are especially important in Cambodia because SMEs “represent 70% of total employment and 99.8% of companies and contribute 58% to gross domestic product (GDP)” according to KhmerSME. A consistent theme in the conference was the need for widespread data availability, sensible AI policies, and foundational AI regulations that can support the growth of SMEs. Open data initiatives can help to improve the availability of data, which is crucial for AI systems. SMEs may be able to use AI to improve the customer experience, optimize supply chains, conduct data-driven decision making, and reach new markets.
The Digital Tech Roadmap, which emphasizes the importance of AI and machine learning, is a sturdy base to further the Cambodia’s digital transformation and economic growth. The Roadmap aims to enhance productivity, efficiency, and innovation for socioeconomic development by leveraging digital literacy, IT skills, and building a robust digital infrastructure. SMEs in Cambodia would benefit from “a legal and regulatory environment that supports the growth of the digital sector and promotes innovation and entrepreneurship while also protecting the rights of consumers and businesses operating in the online space.”
The Cambodian Digital Government Policy 2022-2035 further underscores the role that AI can play in improving public services via digital governance. Southeast Asia scored a 40.5 in Salesforce’s Asia Pacific AI Readiness Index, which puts the region behind many developed East Asia countries like China, Japan, and South Korea. AI systems can benefit from the second level of the Cambodian Digital Government Policy 2022-2035, which seeks to ensure that the “government’s information is open to the public to ensure transparency and accountability.” Open data policies like this can provide high-quality training data for AI systems and enable SMEs to utilize AI systems with contemporary datasets. Enhancing transparency, equality, and accountability via the open data initiative in “Strategy 5: Organizing the Governance of Digital Government” of the Cambodian Digital Government Policy 2022-2035 will cultivate Cambodia’s AI research and innovation.
Several presenters at the conference reiterated the need for a supportive legal, regulatory, and data environment for SMEs to adopt AI and other cutting-edge technologies. The ASEAN Digital Economy Framework Agreement (DEFA) aims to create a standardized regulatory environment that facilitates seamless cross-border data flows and robust digital infrastructure, which are crucial for the development and governance of artificial intelligence (AI) within the region. DEFA’s provisions on AI governance focus on implementing ethical standards, harmonizing AI regulations, and fostering innovation through cooperation among member states. This alignment is expected to enhance knowledge transfer and elevate AI readiness across ASEAN countries, addressing the current disparities in AI adoption and regulatory frameworks.
By promoting interoperability and integration, DEFA supports the establishment of a cohesive digital economy, enabling AI companies to navigate compliance more easily and fostering a conducive business climate for AI-driven innovation and growth. DEFA will drive growth and innovation in the industry while ensuring that the benefits of digital transformation are shared among all stakeholders. Open data policies and robust legal frameworks will support Cambodia’s digital transformation and help progress towards the conference’s central theme of human-centered adoption of AI.
