甫瀚：2026 AI 脉搏调研报告（第四卷）：缺乏 AI 可视性 管控便无底气（英文版）（13页）.pdf

1、A I P U L S E S U RV E Y VO L.4No AI visibility,no confidenceBlind spots dont just hide AI threats they erode confidence in controlsYou cant defend what you cant see;invest in enablers10Defensive AI and training:The ultimate confidence builder09Perception gaps have real consequences05Notable finding

2、s04Why formal frameworks matter07Third-party embedded AI:Where visibility is won or lost08Size doesnt beat blind spots06Survey demographics and methodology11About Protiviti12Executive summary03Table of Executive summary Theres a dual reality playing out in todays cyber landscape:IT leaders see threa

3、ts enabled by artificial intelligence(AI)rising significantly on the front lines,while views from the C-suite are more moderate,according to Protivitis latest AI Pulse Survey,which focuses on cybersecurity and resilience.That same disconnect shows up in leaderships visibility into employees AI tool

4、usage,where we found that many large organisations1(47%)and medium-sized organisations2(68%)face unmanaged AI exposure or shadow AI.3 In essence,boards are funding and governing risk based on an incomplete picture of whats happening across the enterprise.Traditionally,these discrepancies have been a

5、ttributed to the lag that organisations often encounter with new technology adoption and the time it takes for risk outcomes to surface,be measured,and reported up to management and the board.But in todays elevated AI-threat landscape,these delays are serious and warrant careful consideration;they c

6、an translate into slower prioritisation,underfunded controls and governance decisions made on stale signals.Theres encouraging news;most leaders acknowledge that AI is escalating cyber risk and are confident that their security controls are keeping pace with threats.But that is tempered by the fact