The 5 As in AI: A Comparative Review of the EU AI Act and the ASEAN AI Guide
The passage of the EU Artificial Intelligence Act on May 21, 2024, marked a significant milestone in the regulation of artificial intelligence (AI). Designed for EU member states, its implications extend globally. The Act adopts a risk-based approach to AI, categorizing applications based on the perceived threats they may pose to society.
In contrast, the ASEAN AI Guide, released in February 2024, serves as a framework for ASEAN member states, providing guidelines for the ethical development and governance of AI technologies. This article delves into both regulatory frameworks, exploring their similarities, differences, and the potential influence of the EU AI Act on AI governance in the Asia-Pacific region.
1. Autonomy
Both the EU AI Act and the ASEAN AI Guide seek to regulate the extent of autonomy AI systems can operate with, particularly in critical sectors such as healthcare and law enforcement. The EU AI Act emphasizes regulatory compliance, transparency, and accountability, establishing boundaries for autonomous decision-making to prevent unintended consequences in AI deployment.
The ASEAN AI Guide, while also acknowledging the importance of autonomy, presents a non-binding framework aimed at guiding member states in the development of traditional AI technologies.
2. Access
The EU AI Act primarily applies to organizations involved in the design and deployment of AI systems within the EU. This expansive reach includes entities selling or importing AI products for use in the EU. For instance, a US company utilizing AI tools for marketing within the EU would fall under the jurisdiction of the EU AI Act.
Conversely, the ASEAN AI Guide acts as a guidance framework for member states, encouraging compliance with existing laws while promoting harmonization of AI regulations to facilitate cross-border cooperation.
3. Ambit
The EU AI Act introduces a multi-tiered risk classification system, categorizing AI applications based on their potential impact:
- Unacceptable Risk: AI systems that pose significant threats to safety and rights are prohibited, including those that manipulate behavior or involve social scoring.
- High Risk: Systems that could negatively affect fundamental rights must comply with stringent requirements and undergo conformity assessments.
- Limited Risk: These systems face minimal transparency obligations.
- Minimal Risk: Systems that pose negligible risks are not subjected to the EU AI Act but may still be governed by other regulations.
In contrast, the ASEAN AI Guide emphasizes seven guiding principles to foster trust and ethical AI deployment, including transparency, fairness, security, robustness, human-centricity, privacy, and accountability.
4. Accountability
The EU AI Act adopts a risk-based approach for compliance and enforcement, imposing significant penalties for non-compliance, which can range from €7.5 million to €35 million. In contrast, the ASEAN AI Guide encourages voluntary adoption of best practices, lacking specific provisions for accountability.
5. Agility
Both frameworks exhibit a need for agility in response to the rapid evolution of AI technologies. They emphasize the importance of flexibility, allowing businesses to adapt to market changes while balancing the need to protect rights and promote social welfare.
Influence of the EU AI Act on ASEAN Regulations
The EU AI Act has the potential to serve as a blueprint for future AI compliance rules within ASEAN. While the ASEAN AI Guide targets traditional AI applications, it does not encompass generative AI technologies. The clear obligations outlined in the EU AI Act can help businesses understand necessary amendments to their terms and conditions.
As many Southeast Asian nations are still in the early stages of AI governance, the ASEAN AI Guide provides best practices and encourages policy coordination. The ASEAN Digital Masterplan 2025, adopted in 2021, envisions a harmonized digital community, underscoring the significance of cohesive regulations.
Countries like Singapore and Malaysia have already taken steps toward establishing ethical AI frameworks. Singapore’s Model AI Governance Framework and Malaysia’s National AI Framework exemplify regional leadership in responsible AI deployment.
Conclusion
The EU AI Act and ASEAN AI Guide share a commitment to responsible AI use, aligning with societal values and ethical considerations. However, the EU’s stringent regulations present challenges and opportunities for organizations operating within the region. As ASEAN progresses in its regulatory journey, the EU AI Act may serve as a model for future frameworks, fostering a balanced approach to AI governance that addresses both economic benefits and ethical implications.
