Akamai：2025年欺诈与滥用报告：在AI欺诈迷雾中破局寻路（英文版）（48页）.pdf

1、V11 ISSUE 04State of the Internet/SecurityFraud and Abuse Report 2025Charting a Course Through AIs Murky WatersSegmentation Impact StudyA Introduction05 Guest column:Working together to achieve collective defense(Guest contributor:John“JD”Denning)07 Key insights of the report 08 AI bots on the rise1

2、6 A closer look:Infamous bots19 Security spotlight:Evolving bot detection(Guest contributor:David Sncal)20 Industry trends30 Regional trends39 A look at fraud and abuse through OWASP Top 10 lists42Compliance:BalancingregulatorycompliancewithsecurityefficacyinAI-enableddefenses(Guest contributor:Jame

3、s A.Casey)45 Mitigation 46 Conclusion:The evolution of AI bots46 Methodology47 Guest contributors48 CreditsA and Abuse Report 2025IntroductionThe rapid evolution of threat methodologies and a significant uptick in malicious activity prompted the Akamai State of the Internet(SOTI)team to update its a

4、nalysis of the abuse and fraud landscape.In 2025,the level of innovation and complexity in this landscape driven primarily by the proliferation of AI bots and APIs is unprecedented.For clarity,this report defines fraud as the use of deception for unlawful gain,and abuse as the engagement in unauthor

5、ized activities.Although these definitions may appear straightforward,the increasing complexity of AI-powered technologies has blurred the lines between legitimate bot activity and these types of malicious actions.This report focuses on a subset of large language model(LLM)AI bots that we codified a

6、s“Akamai categorized”by using authentication methods that rely on self-identification and bot telemetry.Our research aims to provide organizations with actionable insights into the evolving role of bots and scrapers,to identify the current trajectory of AI-driven threats,and to highlight the threats