当前位置:首页 >英文主页 >中英对照 > 中译版报告详情

波耐蒙研究所:2026 AI、生成式AI与代理型AI的风险管理与价值优化研究报告(中译版)(44页).pdf

上传人: 1****1 编号:1167531 2026-03-26 44页 6.71MB

下载:

1、 Managing Risks and Optimizing the Value of AI,GenAI&Agentic AISponsored by OpenText Independently conducted by Ponemon Institute LLCMarch 2026 Ponemon Institute Research Report Page 1 Managing Risks and Optimizing the Value of AI,GenAI&Agentic AI Sponsored by OpenText March 2026 The purpose of this

2、 research is to gain insight into how organizations are safeguarding their organizations from risks created by AI,GenAI and Agentic AI while still being able to benefit from their use.As shown in this research,organizations believe AI governance is important to realizing AIs value for security and b

3、usiness purposes.AI governance is the process of creating policies,assigning decision rights and ensuring organizational accountability for risks and investment decisions in the application and use of AI technologies.AIs value is dependent upon good governance,non-human identity management and expla

4、inability.Respondents were asked to rate their organizations focus on governance of AI systems from 1=low focus to 10=very high focus.As shown in Figure 1,53 percent of respondents say their organizations are highly or very highly focused on governance(7+responses on the 10-point scale).Non-human id

5、entity management(NHIM)is the process of discovering,securing and managing digital credentials for applications,machines and automated processes that lack a human operator.This practice is considered crucial for cybersecurity because the number of these identities is growing rapidly and they are oft

6、en overlooked,creating significant security vulnerabilities.Respondents were asked to rate the priority of managing non-human identities on a scale from 1=low priority to very high priority.Fifty-one percent of respondents say it is a high or very high priority(7+responses on a 10-point scale).Expla

word格式文档无特别注明外均可编辑修改,预览文件经过压缩,下载原文更清晰!
三个皮匠报告文库所有资源均是客户上传分享,仅供网友学习交流,未经上传用户书面授权,请勿作商用。
1. **AI治理重要性**:53%组织高度关注AI治理,51%认为非人类身份管理(NHIM)是高优先级,47%认为AI可解释性非常重要。 2. **成熟度挑战**:仅21%组织达到AI成熟阶段,50%面临人员不足,46%预算短缺,44%缺乏时间整合AI。 3. **风险与合规**:62%认为降低模型/偏见风险“极其困难”,59%认为数据风险(如低质量训练数据)难以控制,仅41%有AI专用数据隐私政策。 4. **GenAI与Agentic AI**:52%已部署GenAI,但仅45%认为其生成安全洞察高效;38%采用Agentic AI,43%因缺乏风险控制未采用,55%认为AI代理将显著增加数据盗窃风险。 5. **区域差异**:北美(51%)最倾向风险导向治理,亚太(28%)最少采用NIST/ISO框架。
AI治理有多重要? AI风险如何管理? AI价值如何实现?
客服
商务合作
小程序
服务号
折叠