ICTworks

Responsible AI is much more difficult to capture in a broad, catch-all definition, as the very notion of “responsibility” is closely tied to broader digital ecosystems and the social and political contexts in which
AI tools operate and evolve.

Indeed, the concept of “responsibility” varies over time (as AI evolves, so does the idea of how it could/should behave) and across laws, regulations, cultures, and systems of governance. Consequently, a rich but dizzying tapestry of initiatives, measures, and mechanisms has emerged specifically to systemize the responsible and ethical advancement of AI through principles, guidelines, standards, and frameworks aimed at mitigating any potential unintended consequences of AI.

Despite their differences in objective and approach, all these guidelines, standards, and frameworks tend to cover several main themes, suggesting that the conversation around principled AI is beginning to converge, at least among the communities responsible for the development of these documents.

ASEAN Guide on AI Governance and Ethics

The ASEAN Guide on AI Governance and Ethics comprises seven guiding principles that we can use to build trust in AI and ensure AI systems are developed ethically. These principles are largely in line with the various international approaches:

1. Transparency and Explainability: AI actors are responsible for clearly disclosing the implementation of AI solutions to stakeholders and fostering general awareness of the solution. AI actors and organizations deploying AI should also strive to foster general understanding of how AI works with simple explanations on how AI reaches its outcomes.
2. Fairness & Equity: The organization should ensure that algorithmic decisions do not create discriminatory or unjust impacts across different demographics, and the use and development of AI should not result in unfair biasness or discrimination. AI actors should also recognize and make adjustments to rectify imbalances to ensure equity.
3. Security and Safety: AI systems should be safe and secure and not used to deceive or threaten humans and/or human society. A risk-prevention approach should be adopted, and precautions should be put in place so that humans can intervene to prevent harm in the event an AI system makes unsafe decisions.
4. Human-centricity: AI systems should respect human-centered values and pursue benefits for human society, including human beings’ well-being, happiness, etc. Human rights, freedom, and benefits should be respected and protected in the design, development, and deployment of AI systems.
5. Privacy and Data Governance: AI solutions should have proper mechanisms in place to ensure data privacy and protection, and maintain and protect quality and integrity of data throughout their entire lifecycle. Data protocols need to be set up to govern who can access data and when data can be accessed.
6. Accountability and Integrity: AI actors should be accountable for decisions made by AI systems and for their compliance with laws and AI ethics and principles. AI actors should also act with integrity when developing AI systems.
7.  Robustness and Reliability: AI systems should perform consistently as designed, minimize accidental and unexpected hazards, and not be vulnerable to being tampered with or compromised. AI systems should have reproducible results and work reliably for a range of inputs and situations.

The ASEAN Guide on AI Governance and Ethics can help organisations build trust among stakeholders and the public as well as align their AI practices with international standards and best practices, among others. Organisations are encouraged to refer to the guidelines in this document to understand how to assess the risks associated with AI and take measures to design, develop, and deploy AI responsibly in the context of their organisations.

The image above is ChatGPT’s interpretation of a multicultural group of Asian leaders reviewing the ASEAN Guide on AI Governance and Ethics in a spirit of unity and international cooperation with flags of ASEAN countries.