A study commissioned by the Cyber Security Agency of Singapore (CSA) has found that 92% of cybersecurity professionals surveyed in Singapore expect AI-enabled threats to require new skills within the next 12 months. The findings cover changes in cybersecurity work and the capabilities needed to detect and respond to AI-enabled attacks.
Conducted by Blackbox Research and published at Singapore International Cyber Week 2026, Building Singapore’s Cybersecurity Workforce for the Age of AI surveyed 1,000 respondents from government agencies, Critical Information Infrastructure Owners, cybersecurity services companies and the private sector. All respondents were using or piloting AI tools in their work. The report examines changes in cybersecurity roles, skills requirements, hiring needs and career development.
AI tools are changing existing roles
Among those surveyed, 81% of respondents reported that their organisations use AI-enabled cybersecurity tools. Following adoption, more than half described meaningful improvements in the quality, speed and consistency of cybersecurity work.
The study found that 53% reported role augmentation, where professionals perform broadly the same jobs but use AI as part of their work. An equal 53% reported role redesign, referring to situations where AI has significantly changed the skills required for cybersecurity tasks.
Attacker capabilities are another concern, with 87% of respondents believing that attackers are developing their capabilities faster than many organisations in Singapore can defend against. The report identifies detecting and responding to AI-enabled threats, along with securing AI systems themselves, as areas where organisations need to strengthen expertise.
Technical knowledge and human judgement remain essential
Survey respondents are generally positive about what AI could mean for their careers, although they recognise the risk of skills becoming outdated. Specifically, 77% see AI as more of a career opportunity than a threat, while 75% are concerned that parts of their existing skill sets may become less relevant.
For entry-level professionals, strong cybersecurity fundamentals were rated as an important career capability by 65% of respondents. The ability to validate and challenge AI-generated outputs followed at 57%, alongside hands-on technical experience at 51%. These capabilities help professionals understand the systems and threats they encounter, recognise flawed AI responses and develop judgement through experience.
Across all career levels, 69% of respondents rated judgement and problem-solving in business and operational situations highly, making it the highest-rated capability in this group. Cybersecurity domain knowledge and technical fundamentals followed at 66%, with 65% rating the ability to detect and respond to AI-enabled cyber threats highly.
The study emphasises that professionals need to recognise when AI-generated assessments are wrong or incomplete, question their conclusions and determine an appropriate response. That work also involves understanding organisational risk, weighing operational trade-offs and aligning security decisions with business priorities. Respondents identified formal training, practical application and learning on the job as ways to maintain their capabilities.
CSA plans updates to cybersecurity skills frameworks
CSA will use the findings to inform its workforce development work, including a refresh of the cybersecurity track under the Skills Framework for Infocomm Technology and the agency’s Operational Technology Cybersecurity Competency Framework. The aim is to establish a new national cybersecurity competency framework that accounts for emerging technology and changing industry needs.
The planned national framework is intended to be a key component of the Cybersecurity Professional Accreditation Framework. CSA also plans to support employers in redesigning roles and workforce development pathways, educators in keeping training aligned with industry needs, and cybersecurity professionals in continuing to develop the capabilities required for AI-enabled security work.
