新奧爾良警察局通訊部門正在測試由Carbyne公司開發的AI緊急呼叫分類系統，用於應對該市日均超過1000通的緊急電話[1]。該部門於4月已開始利用AI接聽311非緊急呼叫，其中50%的來電涉及信息查詢[1]。
該AI系統主要用於評估和優先處理來電，在呼叫量激增時自動分流同一事件的多重呼叫，但真正的緊急情況仍由人類調度員負責處理[1]。然而，該系統面臨多項潛在風險，包括語音識別失敗、隱含的算法偏見以及可能強化系統性種族主義等問題[1]。
The New Orleans Police Department Communications Division (OPCD) is testing an artificial intelligence system developed by Carbyne to field emergency 911 calls, aimed at managing the volume of incoming requests [1]. The city receives more than 1,000 emergency calls daily [1], and the AI system is designed to assess and prioritize incoming calls while automatically routing multiple reports of the same incident to reduce dispatcher workload [1].
The deployment builds on an earlier initiative launched in April, when OPCD began using AI to answer non-emergency 311 calls, approximately half of which are information requests [1]. Under the new 911 system, artificial intelligence will handle call classification and triage, though human dispatchers will continue to process calls involving genuine emergencies [1].
However, the rollout carries documented risks, including the potential for voice recognition failures, embedded algorithmic bias, and the reinforcement of systemic racial disparities in emergency response [1].