Many organisations in Singapore are currently facing challenges related to AI security, as highlighted by recent research from Proofpoint. The study reveals that 58% of these organisations lack confidence in their AI security systems to detect compromised AI, while half of those with AI security measures reported experiencing AI-related incidents. Notably, 87% have progressed beyond pilot programs with AI assistants, and 70% are in the process of deploying autonomous agents.
The survey, which included over 1,400 security professionals from 12 countries, indicates that AI tools are now a vital part of various business functions, such as customer support and internal communications. However, security governance seems to be lagging behind this rapid adoption, with only 32% of respondents feeling fully prepared to investigate AI-related incidents.
Furthermore, 98% of organisations find it challenging to manage multiple security tools, with 61% rating it as very or extremely difficult. Email is the primary threat vector, identified by 58% of respondents, while many also highlighted risks associated with software-as-a-service applications and collaboration tools like Teams and Slack. The complexity of the attack surface is exacerbated by the interconnectedness of AI systems across multiple platforms.