1、The State of Generative AI20252The State of Generative AI in 2025Table of ContentsExecutive Summary.3Methodology.3Growth Observations:AI Is a Force Multiplier .4GenAI Traffic Increased by More than 890%.4Most GenAI Is Used for Writing,Conversing,and Coding .5Introducing the Top 10 GenAI Apps .6Gramm

2、arly Takes the Top Spot Among Writing Assistantsand All Apps for That Matter.7Microsoft Leads Enterprise Adoption of AI Agents with MS Copilot.8Technology and Manufacturing Use AI Coding Tools a Lot.9GenAI Is Reshaping the SaaS Landscape .11Data and Security Observations:A New Era of Risk.12Shadow A

3、I Is an Emerging Source of Risk .12Expect at Least Six High-Risk GenAI Apps .12GenAI-Related DLP Incidents Are Rising.13Key GenAI Challenges and Risks:AIs Growing Pains .14Navigating the Sensitive Data Minefield .14The Evolving AI Regulatory Landscape .15Staying Ahead of AI Security Blind Spots .15R

4、ecommendations and Best Practices:The AI Game Plan.16Summary:Innovate Fast,but Secure Faster .17Authors.17About Palo Alto Networks .18In our State of Generative AI report,the Palo Alto Networks product and research teams explore the adoption and use of generative AI(GenAI)technology across our custo

5、mers.Our data yielded some noteworthy trends based on the tens of thousands of tenants we observed.Heres what we found:GenAI traffic surged more than 890%in 2024.This explosive growth may be attributable to maturing AI models,greater enterprise automation,and higher adoption due to more evident prod

6、uctivity gains.Increased adoption and usage marks a definitive shift from GenAI as a novelty to an essential utility.Data loss prevention(DLP)incidents for GenAI more than doubled.In 2025,the average monthly number of GenAI-related data security incidents increased 2.5X,now comprising 14%of all data