Governing the risks of advanced artificial intelligence (AI) requires a multi-step, “whole-of-organization” approach to avoid the Philippines from being overwhelmed by external, AI-related threats, according to a legal expert.
The Philippines ranks 67th out of 83 countries in the Global AI Index, with an overall score of 5.9 out of 100 in AI development, placing it in the bottom quarter just above Rwanda and below Slovakia.
Meanwhile, global AI risks and harms continue to rise, as reported by the Organisation for Economic Co-operation and Development (OECD) in its Incidents and Hazards Monitor.
Edsel Tupaz, a senior partner at Gorriceta Law and Head of the firm’s Data Privacy, Cybersecurity, and AI Initiatives, said addressing these risks must be “across the board” as AI is a “multidisciplinary, cross-domain, and cross-functional” discipline.
Tupaz proposed a multi-step governance strategy for organizations to address these risks through legal, regulatory, and governance frameworks. The approach starts with ensuring compliance with sector-specific regulations and industry requirements.
Next comes training, oversight and accountability. Organizations should “build internal capacity through training and awareness, and also locate accountability among the owners of AI processes,” he said.
The following step is continuous optimization of policies, processes, and controls through regular assessments and audits.
“In other words, this is really the norm of continuous improvement under ISO standards,” Tupaz said.
“Audits and regular assessment conduct what we call AI impact assessments in addition to your privacy impact assessment. You can deploy AI governance and risk management systems, such as ISO and NIST,” he said.
He also cited the need for cybersecurity testing, particularly “red teaming,” where organizations take an “adversarial lens in testing [their] systems to see if they can be robust and withstand potential adversarial attack.”
Despite being a large and attractive market for foreign AI providers, the Philippines remains at the “tail end” of the global AI supply chain, serving primarily as an importer and end-user.
“I think the posture there should be actors within the Philippines, organizations, companies, firms ought to consider recasting our role as an active governance actor,” he said.
“Not passive, not waiting for some large solution, but an active governance actor who can better assert agency in the name of AI safety,” he said.
