Discussion Paper on Securing Artificial Intelligence (AI): A Collective Responsibility

Read about Securing AI at Issuu.

AI has the potential to drive efficiency and foster innovation across sectors, from healthcare and e-commerce, to the delivery of government services. In cybersecurity, it is expected to alleviate the workload for analysts and operators while aiding in the combat emerging threats.

However, the adoption of both traditional and generative AI will exacerbate existing security risks and introduce new challenges. As with existing digital systems, malicious attacks on AI can undermine user confidence in the technology, and ultimately limit the value they can derive from it.

This discussion paper by the Cyber Security Agency of Singapore and Resaro explores the intricacies of AI security and its relation to current international conversations about the safety and trustworthiness of AI. The paper also examines how AI security differs from traditional IT security, and the unique risks and vulnerabilities that arise from inherently adaptive nature of AI. The paper then explores how these risks can be addressed through a collaborative, multi-stakeholder effort at both the organisational and ecosystem levels, as well as discuss the roles that various players in the AI ecosystem should undertake to ensure secure AI adoption.

To advance AI security, the paper proposes four key pillars that governments, industry, experts, and academia to focus on:

(a) building robust and practical AI governance frameworks;
(b) developing technical tools and capabilities that can support secure AI adoption;
(c) investing in research and development to identify emerging threats to AI security and mitigation strategies; and
(d) supporting talent, innovation, and growth (TIG) programmes that foster the growth and development of a secure AI ecosystem.

• Securing AI: A Collective Responsibility [PDF, 638 KB]