Comparative Analysis of ASEAN AI Guide and EU AI Act
Amy Darajati Utomo is an international development practitioner with a diverse background spanning five years across government, the private sector, and nonprofits. Recently completing the MPA in Development Practice program at Columbia University, her passion lies in leveraging technology to advance gender equality. Amy’s previous roles at the ASEAN Foundation, the World Bank Group, and the United Nations have equipped me with valuable insights into utilizing technology for inclusive development.
Currently, Amy serves as the program manager at iamtheCODE Foundation, where she focuses on advancing digital education for marginalized communities. This role has deepened her commitment to fostering an equitable digital world, where technology serves as a catalyst for positive social change.
Abstract: This paper undertakes a comparative analysis of the regulatory approaches to Artificial Intelligence (AI) governance adopted by the Association of Southeast Asian Nations (ASEAN) and the European Union (EU). Specifically, it examines the ASEAN Guide on AI Governance and Ethics alongside the EU's AI Act. These two regions present contrasting regulatory philosophies, offering insights into how they navigate the balance between fostering innovation and ensuring ethical AI use.
While both frameworks aim to achieve this balance, they diverge significantly in their methods and underlying principles. The differences stem from various factors, including distinct legal and political structures, economic contexts, and philosophical approaches to AI governance. ASEAN and the EU prioritize different aspects of AI regulation based on their regional priorities and challenges.
Despite these disparities, the paper suggests that harmonization between the ASEAN AI Guide and the EU AI Act is feasible. Collaborative efforts between ASEAN and the EU could establish global standards that harmonize innovation with ethical considerations in AI development and deployment. Such collaboration could pave the way for a more cohesive global AI governance framework, ensuring consistency and effectiveness across international borders.
As the capabilities of artificial intelligence (AI) evolve, governments worldwide are actively engaged in discussions to ensure the ethical and responsible use of AI technologies. These dialogues extend beyond domestic legislation, with noticeable endeavors to regulate AI emerging within numerous regional organizations. It is intriguing to observe the collaborative efforts among diverse nations in crafting regulatory frameworks in response to AI, and to discern the variations that may arise across different regions.
The ASEAN Guide on AI Governance and Ethics (hereafter referred to ASEAN AI Guide) and the European Union Artificial Intelligence Act (hereafter referred to EU AI Act) represent two distinct approaches to AI regulation that reflect the diverse regulatory philosophies of their respective regions. This paper will compare the ASEAN AI Guide's light touch, risk-based approach with the EU's top down, hard regulatory AI Act.[1] Furthermore, it will also explore how these regulatory frameworks could influence the global AI governance landscape.
ASEAN AI Governance Guide: A Flexible and Business-Friendly Approach
The Association of Southeast Asian Nations (ASEAN) published the ASEAN AI Guide in February 2024, providing regional guidance for the ethical development and deployment of AI technologies. ASEAN holds enormous growth potential in the digital economy and the possibility of integrating its manufacturing and creative economy sectors with digitalization. A report by Google, Temasek and Bain & Company shows that Southeast Asia’s digital economies are expected to reach $218 billion in total value of transactions in 2023, an increase of 11% from 2022, amid global downturns.[2]
The guide is voluntary and encourages innovation by allowing companies to adapt the principles to their specific contexts. The guide's seven principles—transparency, fairness, security & safety, human-centricity, privacy, accountability, and robustness—are meant to be translated into business practices.[3] It includes AI governance use cases from companies and the public sector in ASEAN. The annexes provide six case studies on AI use as well as a 11-page risk impact assessment with example questions and useful industry examples, practices and guides for consideration. The intention is to help promote consumer confidence and facilitate cross-border deployment of AI-powered services and solutions, enabling ASEAN to fully leverage the power of AI as a region. The Guide was developed in consultation with industry partners, and was endorsed in February 2024.[4]
EU AI Act: Striving for Harmonization and Stringent Regulation
The EU AI Act is the world’s first sweeping AI law[5] that regulates the development, deployment, and use of AI systems within the EU. The Act, which was originally proposed in April 2021 and reached a provisional agreement in December 2023, and on March 13, 2024, the European Parliament voted to approve the EU AI Act. It aims to ensure that AI systems used in the EU are transparent, reliable, and safe, while also respecting fundamental rights and values. It does so through a combination of prohibitions, mandatory requirements, transparency obligations, conformity assessments, and a governance framework, all based on the level of risk posed by specific AI systems.
Figure 1 - EU AI Act's Risk-Based Approach [6]
The EU AI Act classifies AI systems into four risk levels:[7]
- Unacceptable risk: AI applications that pose an unacceptable risk, such as government-operated social scoring and manipulative AI techniques, are outright prohibited.
- High risk: AI systems with high risk, like CV-scanning tools for employment purposes, are subject to stringent requirements. These requirements cover areas such as risk management, data governance, transparency, human oversight, accuracy, robustness, and cybersecurity. High-risk AI systems must undergo conformity assessments before being deployed or made available on the market.
- Limited risk: AI systems with limited risk, such as chatbots, are obligated to be transparent. For example, chatbots must disclose to users that they are interacting with an AI.
- Minimal or no risk: AI systems with minimal or no risk are not subject to regulation.
The Act establishes the European AI Board and national competent authorities responsible for enforcement. Penalties for non-compliance can result in fines of up to €30 million or 6% of global annual turnover. The AI Act applies to providers and users of AI systems in the EU, as well as providers outside the EU if their systems are used within the EU.
Comparing ASEAN AI Guide and EU AI Act
The ASEAN and EU approaches to AI regulation offer contrasting perspectives, which can be summarized in the table below.
Based on the table above, it is clear that while both frameworks aim to balance innovation with ethical considerations, their methods differ significantly. ASEAN's flexible approach allows for rapid adaptation and innovation, potentially making it easier for emerging AI technologies to reach the market. Conversely, the EU's structured approach, while potentially slower in accommodating new technologies, emphasizes robust ethical considerations and consumer protections.
Key Factors Contributing to the Differences
Several key factors contribute to the differences between the EU AI Act and the ASEAN AI Guide. Firstly, legal and political structures differ. The EU is a supranational organization, where member states have pooled their sovereignty in certain areas, such as trade and environment. The EU has a centralized legislative body that can pass binding laws across all member states, such as the AI Act. They have institutions such as a Parliament, a court, a single currency and a foreign service which ASEAN does not have. However, the ASEAN, an intergovernmental organization, lacks a similar region-wide legislative body. Its policies, like the AI Guide, are voluntary and meant to guide domestic regulations. The ASEAN Secretariat, with fewer than 400 staff, is tiny compared with the European Commission (EC) with more than 23,000 personnel. ASEAN's annual budget of about US$20 million (S$27 million) is minuscule compared with that of the EU Commission, with a budget of about US$180 billion.[8]
Secondly, differences extend to their economic and political contexts, with ASEAN's diverse landscape requiring an adaptable framework. With disparate rules on issues like censorship and intellectual property, this diversity complicates efforts to consensus building, for example, to agree on how to address potentially harmful AI-generated content. Furthermore, ASEAN encompasses a spectrum of economies, spanning from advanced nations like Singapore, which boast well-established and flourishing tech industries, to countries where internet access is a recent development for much of the population, accompanied by low levels of digital literacy. Meanwhile, The EU focuses more on harmonizing rules and values across member states, and the EU has progressed much further with its single market and institutions. [9] The EU's more uniform political-economic landscape and pooled sovereignty enable it to enact comprehensive, standardized frameworks to address digital divides.[10]
Thirdly, the regions have different philosophical approaches. The EU AI Act reflects a precautionary approach, strictly regulating AI systems based on their risk levels to prevent potential harms. The ASEAN AI Guide adopts a more permissive stance, providing guidelines to manage AI risks while allowing flexibility for countries to adapt them to their contexts. This flexibility allows consideration of cultural differences, something that the EU AI Act does not address explicitly. ASEAN AI Guide advises business and tech players to take countries’ cultural differences into account as ASEAN is a region that enjoys cultural and linguistic diversity across countries and communities.[11]
Fourthly, the difference shows the variance of priorities in balancing innovation and regulation. ASEAN countries, particularly those with developing digital infrastructures, are wary of over-regulating AI, fearing it might constrain innovation or drive it elsewhere.[12] Therefore, it is very business friendly, as it was developed through consultation with industry partners. The Guide also stated that it “is meant to be a living document that should be periodically reviewed and assessed by relevant ASEAN sectorial bodies, in consultation with industry partners.” Therefore, ASEAN’s decision to shy away from the EU’s strict AI guidance highlights its regional ambition to advance its economy through technological innovation.[13]
Meanwhile, the EU represents a more precautionary approach to AI governance, prioritizes stricter rules to protect rights and mitigate AI risks.[14] The EU's approach has faced pushback from the business community, with executives warning that the legislation could jeopardize Europe's competitiveness, investment, and innovation. [15]
Finally, the regions also differ in terms of their goal for their framework. The EU aims to set global standards for AI governance aligned with its values and is actively promoting its AI Act to other regions.[16] EU officials were sent to a dozen Asian countries, including Singapore and the Philippines, last summer to convince national governments to back its more stringent AI rules. Meanwhile, ASEAN countries are cautious about adopting the EU's stringent approach wholesale, preferring a framework that suits their regional priorities and development goals. Ironically, EU member states formally backed the AI Act at a meeting on February 2, the same day that ASEAN released its guidelines.[17] It suggests that ASEAN is steering clear of aligning with any particular global power in the geopolitical race for AI advancement and business opportunities.[18]
Despite these differences, harmonization between the ASEAN AI Guide and the EU AI Act is possible. Both regions could collaborate to establish global standards that balance innovation with ethical considerations, leading to a more cohesive global AI governance framework.
Conclusion
In conclusion, the ASEAN AI Guide and the EU AI Act represent two ends of the spectrum in AI governance. ASEAN's guide promotes a flexible, business-friendly approach that encourages innovation, while the EU AI Act leans towards a more regulated environment that prioritizes societal protection. The global AI governance landscape could benefit from a synthesis of these approaches, where innovation is fostered without compromising ethical standards and societal well-being. As the global community continues to grapple with the challenges of AI, the potential for harmonization and mutual learning between these frameworks offers a path towards a balanced and effective governance of AI technologies.
Bibliography
"Artificial Intelligence Act." Accessed May 2024. [https://artificialintelligenceact.eu/]
Baijal, Aadarsh, Florian Hoppe, Willy Chang, Sapna Chadna, and Fock Wai Hoong. "E-Conomy SEA 2023." Accessed May 2024. [https://www.bain.com/insights/e-conomy-sea-2023/]
Crowden, Maya, and Mario Masaya. "ASEAN Releases Regional Guide on AI Governance and Ethics." Accessed May 2024. [https://www.usasean.org/article/asean-releases-regional-guide-ai-governance-and-ethics]
Edwards, Lilian. "EU AI Act Explainer." Accessed May 2024. [https://www.adalovelaceinstitute.org/resource/eu-ai-act-explainer/]
Elbashir, Mohamed. "EU AI Act Sets the Stage for Global AI Governance: Implications for US Companies and Policymakers." Accessed May 2024. [https://www.atlanticcouncil.org/blogs/geotech-cues/eu-ai-act-sets-the-stage-for-global-ai-governance-implications-for-us-companies-and-policymakers/]
European Union. "Key Facts and Figures - Economy." Accessed May 2024. [https://european-union.europa.eu/principles-countries-history/key-facts-and-figures/economy_en]
Hutt, David. "AI regulations: What can the EU learn from Asia?" Accessed May 2024. [https://www.dw.com/en/ai-regulations-what-can-the-eu-learn-from-asia/a-68203709]
Ministry of Foreign Affairs - Permanent Mission of the Republic of Singapore. "ASEAN and EU: The untold story." Accessed May 2024. [https://www.mfa.gov.sg/Overseas-Mission/Ministry-of-Foreign-Affairs---Permanent-Mission-of-the-Republic-of-Singapore/Recent-Highlights/2020/10/ASEAN-and-EU-The-untold-story]
Personal Data Protection Commission Singapore. "Two Guides Unveiled at ADGMIN 2024." Accessed May 2024. [https://www.pdpc.gov.sg/news-and-events/announcements/2024/02/two-guides-unveiled-at-adgmin-2024/]
Toh, Michelle. "EU companies face fines up to 6% of global revenue under proposed AI law." Accessed May 2024. [https://www.cnn.com/2023/06/30/tech/eu-companies-risks-ai-law-intl-hnk/index.html]
Wang, Grace Yuehan. "Southeast Asia Artificial Intelligence Governance: Guide in Formation, Picking No Side Among Global Powers." Accessed May 2024. [https://blogs.lse.ac.uk/internationaldevelopment/2024/02/26/southeast-asia-artificial-intelligence-governance-guide-in-formation-picking-no-side-among-global-powers/]
"European approach to Artificial Intelligence." Accessed May 2024. [https://digital-strategy.ec.europa.eu/en/policies/european-approach-artificial-intelligence]
"High Level Summary." Accessed May 2024. [https://artificialintelligenceact.eu/high-level-summary/]
ENDNOTES
[1] Maya Crowden and Mario Masaya, "ASEAN Releases Regional Guide on AI Governance and Ethics," US-ASEAN Business Council, accessed May 2024, https://www.usasean.org/article/asean-releases-regional-guide-ai-governance-and-ethics.
[2] Florian Hoppe, Aadarsh Baijal, Willy Chang, Sapna Chadna, and Fock Wai Hoong, "E-Conomy SEA 2023," Bain & Company, accessed May 2024, https://www.bain.com/insights/e-conomy-sea-2023/.
[3] Association of Southeast Asian Nations (ASEAN), "ASEAN Guide on AI Governance and Ethics" (Version 2, February 2024), accessed May 2024, https://asean.org/wp-content/uploads/2024/02/ASEAN-Guide-on-AI-Governance-and-Ethics_beautified_201223_v2.pdf.
[4] Personal Data Protection Commission Singapore, "Two Guides Unveiled at ADGMIN 2024," accessed May 2024, https://www.pdpc.gov.sg/news-and-events/announcements/2024/02/two-guides-unveiled-at-adgmin-2024.
[5] "Artificial Intelligence Act," accessed May 2024, https://artificialintelligenceact.eu/.
[6] Lilian Edwards, "EU AI Act Explainer," Ada Lovelace Institute, accessed May 2024, https://www.adalovelaceinstitute.org/resource/eu-ai-act-explainer/.
[7] "High Level Summary," accessed May 2024, https://artificialintelligenceact.eu/high-level-summary/.
[8] Ministry of Foreign Affairs - Permanent Mission of the Republic of Singapore, "ASEAN and EU: The untold story," accessed May 2024, https://www.mfa.gov.sg/Overseas-Mission/Ministry-of-Foreign-Affairs---Permanent-Mission-of-the-Republic-of-Singapore/Recent-Highlights/2020/10/ASEAN-and-EU-The-untold-story.
[9] Ministry of Foreign Affairs - Permanent Mission of the Republic of Singapore, "ASEAN and EU: The untold story,"
[10] European Union, "Key Facts and Figures - Economy," accessed May 2024, https://european-union.europa.eu/principles-countries-history/key-facts-and-figures/economy_en.
[11] Grace Yuehan Wang, "Southeast Asia Artificial Intelligence Governance: Guide in Formation, Picking No Side Among Global Powers," LSE International Development Blog, February 26, 2024, accessed May 2024, https://blogs.lse.ac.uk/internationaldevelopment/2024/02/26/southeast-asia-artificial-intelligence-governance-guide-in-formation-picking-no-side-among-global-powers/.
[12] David Hutt, "AI regulations: What can the EU learn from Asia?," accessed May 2024, https://www.dw.com/en/ai-regulations-what-can-the-eu-learn-from-asia/a-68203709.
[13] Grace Yuehan Wang, "Southeast Asia Artificial Intelligence Governance: Guide in Formation, Picking No Side Among Global Powers,"
[14] Mohamed Elbashir, "EU AI Act Sets the Stage for Global AI Governance: Implications for US Companies and Policymakers," Atlantic Council, accessed May 2024, https://www.atlanticcouncil.org/blogs/geotech-cues/eu-ai-act-sets-the-stage-for-global-ai-governance-implications-for-us-companies-and-policymakers/.
[15] Michelle Toh, "EU companies face fines up to 6% of global revenue under proposed AI law," CNN, June 30, 2023, accessed May 2024, https://www.cnn.com/2023/06/30/tech/eu-companies-risks-ai-law-intl-hnk/index.html.
[16] "European approach to Artificial Intelligence," European Commission, accessed May 2024, https://digital-strategy.ec.europa.eu/en/policies/european-approach-artificial-intelligence.
[17] David Hutt, "AI regulations: What can the EU learn from Asia?," DW.com, accessed May 2024, https://www.dw.com/en/ai-regulations-what-can-the-eu-learn-from-asia/a-68203709.
[18] Grace Yuehan Wang, "Southeast Asia Artificial Intelligence Governance: Guide in Formation, Picking No Side Among Global Powers,"
