Arkose Labs：2025年网络安全领域人工智能（AI）成熟度报告（英文版）（64页）.pdf

1、Can Defenses Catch Up Before Agentic AI Catches On?Enterprises doubled their AI security budgets and confidence in just 12 months.But 53%expect agentic AI attacks to hit within six months.When threats evolve faster than implementation cycles,who wins the race?The Arkose Labs 2025 AI Maturity in Cybe

2、rsecurity ReportTABLE OF CONTENTS2Executive Summary03Key Findings04The Cost of Being a Target05The Investment Surge10The Mobilization Year15The Agentic AI Reckoning20Know Your(AI)Agent27About Arkose Labs62Appendix63Recommendations60Methodology61Industry AI Scorecards Airline Banking Fintech Hotel Sh

3、aring/Gig Economy Social Media Streaming Media Technology36EXECUTIVE SUMMARYSecurity teams have six months,or less,to prepare for agentic AI attacks.Implementation cycles take 12-18 months.The math doesnt work.That timeline gap drives everything revealed in Arkose Labs second annual AI maturity surv

4、ey:a cybersecurity landscape where speed determines survival.Enterprises nearly doubled their very well prepared ratings from 23%to 44%year-over-year and committed 32%of cybersecurity budgets to AI solutions,up from 27%in 2024.The buy-versus-build debate?Settled.Eighty-one percent now gain more valu

5、e from purchasing solutions than building them,up 19 points from 62%as vendor partnerships evolved into operational extensions of security teams.Integration barriers dropped 11 points,removing what was previously the primary obstacle to effectiveness.Can confidence substitute for capability?The data

6、 says no.Its an arms race where adversaries adapt and improvise as fast as defenders deploy.Fifty-five percent still lose$10 million-$500+million annually.Over half of enterprises already use agentic AI for cybersecurity,but 71%see bad actors use of agentic AI as the next major threatone that 53%exp