ASEAN issues AI guidelines

Framework set to aid responsible and secure use of new tech within region

“Misinformation and disinformation can be supercharged with deepfakes generated by AI. All governments will be challenged to ensure that digital developments are built upon a strong foundation of trust.”

JOSEPHINE TEO Communications and information minister of Singapore

The publication of a guide for artificial intelligence (AI) governance by the Association of Southeast Asian Nations (ASEAN) in line with the group’s goal to establish policies and guidance to promote technologies in a “responsible and secure manner” can contribute significantly to global AI discussion, analysts say.

The two-day 4th ASEAN Digital Ministers Meeting in Singapore concluded on Feb 2 with the unveiling of the ASEAN Guide on AI Governance and Ethics, saying it will serve as a “practical and implementable tool” to support the deployment of AI solutions in the region.

The guide includes national and regional recommendations that organizations in member countries can consider applying in designing, developing, and deploying AI systems in their diverse and tech-savvy populations.

ASEAN includes Brunei Darussalam, Burma, Cambodia, Indonesia, Laos, Malaysia, Philippines, Singapore, Thailand, and Vietnam.

Josephine Teo, communications and information minister of Singapore and rotating chair of the ministerial meeting, said the current wave of digital technology has the potential to sow distrust.

“Misinformation and disinformation can be supercharged with deepfakes generated by AI. All governments will be challenged to ensure that digital developments are built upon a strong foundation of trust,” Teo said.

The guide arrives amid global debate about how to set up guardrails to ensure that this emerging technology remains safe and ethical.

Elina Noor, a senior fellow in the Asia Program at the Carnegie Endowment for International Peace, said ASEAN needs to be bolder in its contributions to global AI discussions, with proposals based on the region’s experience with technology.

“We tend to consider AI in a time bubble of the present and the future without a fuller reflection of how the fundamental basis of AI, which is data, has shaped and will continue to shape different aspects of our society,” Noor said.

One of the regional recommendations is to establish an ASEAN working group on AI governance that will oversee AI governance initiatives in the region.

Kan Min Yen, an associate professor at the National University of Singapore’s School of Computing, said ASEAN is a “critical region” for AI governance.

“Current AI systems are trained on data mostly unrepresentative of ASEAN citizens,” Kan said, adding that ASEAN’s local languages, cultural norms, and even everyday objects are not well represented in data used for AI system training.

The guide cites seven guiding principles: transparency and explainability, fairness and equity, security and safety, human-centricity, privacy and data governance, accountability and integrity, and robustness and reliability.

Dominic Ligot, executive director of Data Ethics PH in Manila, said ASEAN can play a big role in crossborder dialogue on emerging best practices in AI.

Apart from launching a guide on AI, the ASEAN ministers reaffirmed the importance of a secure, diverse and resilient digital infrastructure within the region. They also said that they plan to deepen the cooperation with dialogue and development partners in enhancing digital and cybersecurity cooperation.

Separately, the 27 member states of the European Union on Feb 2 unanimously endorsed the bloc’s landmark legislation governing the utilization of artificial intelligence (AI), also known as the EU AI Act.

European Commissioner for the Internal Market Thierry Breton described the move as “historic”, emphasizing that negotiators have found the “perfect balance between innovation and safety”.

In December, EU lawmakers and member states reached political agreement on the key provisions of the AI Act, hailed by the European Commission as the “first-ever comprehensive legal framework on AI worldwide”.

However, the complexity of the legislation prompted monthslong technical refinement and a new round of discussions.

The European Parliament is anticipated to vote on the finalized text in March or April before it becomes law.

While the legislation will not take immediate effect, certain rules will be implemented within six months after being gazetted.