安全领域的人工智能：两年经验——哪些有效哪些无效.pdf

1、CHATGPT IN YOUR SECURITY PROGRAMTwo Years In-What Worked,What DidntSHAWN CHAKRAVARTYPronounced Chuck Ruh Var TeeGSLC,GCTI,GISP,GSOM,GCIL,AWS Solutions ArchitectMini-ResumeCertified SANS Instructor LDR512Upwork Sr.Director of Active DefenseBuilding Security Operations Centers for large enterpriseCybe

2、r Security 14 years(2012)NOC and IT since 1998Amateur Blacksmith and knife maker(Forged in Fire S7 E9)WHERE WE STARTED“We are all in on AI”-LeadershipBuilding an AI policyDebating with the GRC and Legal teamsBuilding a realistic road map for harnessing AIWhat you build can benefit teams outside of s

3、ecurityI WAS PROMISED A FLYING CARNo more alert fatigueForget about having to train your analystsIn 3 years your L1s will only be doing L2 workThink of all the people hours you will save.Phishing will be a thing of the pastAutomated IRThreat intelligence catered to your businessWHAT WAS PREDICTED AI

4、 to generate SIEM queries AI to handle all L1 triage AI to recommend code improvements AI to prioritize vulnerabilities AI to build automation AI to do phishing email analysisWHAT WE GOT AI to generate SIEM queries Proved Useless AI to handle all L1 triage Kind of AI to recommend code improvements O

5、k this is actually awesome AI to prioritize vulnerabilities It is getting there AI to build automation Not that I have seen AI to do phishing email analysis A give and takeWHERE TO START Make sure your company has an AI Policy.What do your people hate doing?What do they complain about?When you ask s

6、omeone to write something what takes the longest to complete?Where can you set an expectation with the target audience that will curb the most risk?Where are the gaps/what are you not able to focus on?TOOLS TO CONSIDERChatGPT EnterpriseGemini EnterpriseCursorGombocClaudeInternally developed LLMsWHAT