RSAC forecasts AI agents will reshape enterprise security in 2027
According to iThome Taiwan, RSA Conference research forecasts that AI agents will discover and patch more vulnerabilities in 2027, forcing security leaders to rethink vulnerability management, staffing, GRC, incident response, and recovery. RSAC also expects more stateless systems to move to the cloud, where clean images can replace compromised instances, while stateful systems still require isolation and patching. Its research additionally identifies MCP agent security, model-generated tool calls, model poisoning, and malicious package recommendations as emerging risks.
The forecast indicates that enterprise security teams may need to redesign both staffing and remediation workflows around AI agents, while also tightening permissions for tool integrations and reviewing dependencies introduced by AI coding assistants.
資安產業會議RSA Conference(RSAC)從2026年大會議程及參與情況觀察,預測2027年資安主管與第一線人員的關注重心將有所不同。AI代理將發現及修補更多漏洞,企業資安團隊也必須重新檢視漏洞管理流程、資安人力及系統復原能力。
資安主管重新聚焦GRC,第一線人員加強事件應變
RSAC在2026年6月發布的研究報告指出,AI與機器學習安全(AI & ML Security)首次成為2026年RSAC大會最受關注的議題。不過,到了2027年,資安主管在大會議程中的關注重心預期將重新轉向GRC,第一線資安人員則會增加對事件應變及復原的投入。
這代表企業管理階層須處理AI治理、法規遵循及風險管理,第一線資安團隊則須調整既有漏洞處理流程,以因應工作量增加。
雲端安全也將重新受到第一線資安人員重視。RSAC預測,企業可能將更多無狀態系統(Stateless Systems)移轉至雲端,系統遭到入侵或需要修補時,便可透過乾淨的系統映像檔重新建立執行個體,減少逐一修復原有系統的作業。