Managing AI Agents Like New Hires Raises Accountability Concerns
TechNews reports that Dot Compliance CEO Doron Sitbon recommends treating enterprise AI agents like new hires, with defined roles, high-value pilot deployments, style guides, human escalation paths, performance reviews, and manager training. A Boston University and BCG Henderson Institute experiment involving 1,261 managers found that calling an agent an “AI employee” reduced managers’ error detection by 16% and weakened their sense of responsibility for the output. The analysis says the new-hire metaphor may aid onboarding and oversight, but final accountability should remain with humans.
The report pairs practical enterprise deployment guidance with experimental evidence that employee-style framing can reduce both error detection and managerial accountability, making oversight design central to AI-agent adoption.
AI 代理（AI agent）快速進入企業工作流程，但許多組織仍把它當成「裝好就該照指令運作」的軟體工具，一旦表現不如預期，員工便心生挫折，導入計畫還沒啟動就先卡關。生命科學品質與法規遵循解決方案商 Dot Compliance 執行長 Doron Sitbon 在《富比士》（Forbes）撰文指出，與其把 AI 視為工具，不如把它當成新進員工看待。
Sitbon 認為，挫折感多半來自期待落差，而非技術缺陷。企業不會要求新人第一天就熟悉所有流程，而是給予學習時間、介紹給團隊，並隨成長調整期待，AI 代理也應比照辦理。面對員工擔心被取代的焦慮，他建議企業以「新同事報到」方式溝通，說明代理人將接手哪些工作、哪些事情會改變，以及設有哪些防護機制，讓焦點回到工作本身而非技術。
#Dot Compliance#Doron Sitbon#Boston University#BCG Henderson Institute#NTT DATA#Atlassian#Cisco#Workday#Gartner#Cloud Security Alliance#Observatorio de RR.HH.