社團指南

歐盟禁止工作場所與學校情緒辨識

The EU AI Act prohibits using AI to infer emotions of natural persons in workplace and education settings, except when the system is intended for medical or safety reasons.

聆聽此指南AI 語音敘述 · ~3 min

本頁閱讀時間3分鐘
  1. 概述
  2. 深入探討
  3. 現實世界的實施
  4. The Future of EU Ban on Emotion Recognition at Work and in Schools
  5. 付諸實踐
  6. 不斷探索
  7. 常見問題

概述

The prohibition has applied since 2 February 2025 and is distinct from high-risk obligations.

深入探討

Article 5(1)(f) prohibits placing on the market, putting into service for this specific purpose, or using AI systems to infer emotions of natural persons in workplace and education institutions. The exception is for systems intended to be put in place or on the market for medical or safety reasons. It is a purpose-specific exception, not a general permission for employers or schools to monitor mood. The Act defines an emotion recognition system as an AI system identifying or inferring emotions or intentions from biometric data. The prohibition does not cover physical states such as pain or fatigue under the statutory definition, and it does not treat detection of readily apparent expressions, gestures, or movements as emotion recognition merely because those signals correlate with feelings. But a vendor’s claim that it measures engagement, attention, enthusiasm, or frustration should be examined by what the system actually infers from biometric data and how it is used. Renaming an emotion score does not decide legal scope. This is a prohibited practice, not a high-risk classification that can be satisfied through documentation, accuracy, or human review. A workplace medical or safety purpose may fall within the exception, but teams need a real, documented fit between that purpose and the system. A general productivity or performance goal is not transformed into a safety purpose by assertion. Other laws, including data protection and employment protections, remain relevant even if an exception applies. The prohibition entered into application on 2 February 2025. Employers, schools, and providers should identify biometric inputs, intended outputs, purpose, affected people, and any onward decisions. If the system infers emotional state in a covered setting, stop and obtain qualified legal review before deployment. Do not collect biometric data simply to test whether a prohibited feature might be useful.

現實世界的實施

A school rejects a webcam tool that labels students as bored based on facial movement.

An employer checks whether a “wellbeing” dashboard infers workers’ emotional states from voice or face data.

A safety team evaluates a fatigue alert for a driver and documents the narrow safety purpose and applicable safeguards.

A buyer asks a vendor what its engagement score actually infers before approving procurement.

The Future of EU Ban on Emotion Recognition at Work and in Schools

The Commission’s current AI Act page lists workplace and education emotion recognition among prohibited practices and confirms these prohibitions applied from February 2025. Interpretive guidance may clarify borderline systems, while courts and national authorities assess concrete deployments. Organizations should keep a dated analysis of system function, setting, and purpose and revisit it when vendors change models or add new inferences. Medical and safety exceptions require careful assessment under the law and do not displace other safeguards. Repeat the scope review if a system adds biometric inputs, new emotion labels, or a new decision pathway.

付諸實踐

為什麼這很重要

  • 風險與安全. 災難性和日常的人工智慧危害都取決於誰了解風險以及誰能夠採取行動。
  • 更明確的決策. 民眾和專業素養決定強而有力的安全政策在政治上是否可行。
  • 突破炒作. 清晰的解釋可以減少炒作、實驗室公關和模糊道德劇場的影響。

留意

  • 將存在風險視為科幻小說,同時能力複合。
  • 混淆了表面產品安全與高度自治下的對準。
  • 只給非英語和非專業觀眾留下低品質的資源。

後續步驟

  1. 單獨的產品危害、誤用和失控/失調風險。
  2. 詢問哪些證據會改變您對時間表和嚴重性的看法。
  3. 比起行銷主張,更喜歡主要來源和具體評估。
  4. 確定一條行動路徑:職業、政策、資金或技能——而不僅僅是意識。

不斷探索

免費電子報

獲取每日 AI 簡報

每個工作日上午提供三則經過驗證的 AI 新聞,以淺顯易懂的英文撰寫。永久免費,無廣告。

每個工作日一封電子郵件。一鍵取消訂閱。我們絕不出售或分享您的地址。

測試一下自己

Take the EU Ban on Emotion Recognition at Work and in Schools quiz

每個答案皆有即時回饋,通過課程後可獲得具備可驗證 ID 的可分享證書。

開始測驗

支持免費的 AI 教育。AI Understanding 是一個 501(c)(3) 非營利組織——絕無廣告,絕無付費牆。 捐款

常見問題

What is EU Ban on Emotion Recognition at Work and in Schools?

The EU AI Act prohibits using AI to infer emotions of natural persons in workplace and education settings, except when the system is intended for medical or safety reasons. The prohibition has applied since 2 February 2025 and is distinct from high-risk obligations.

Which use does Article 5(1)(f) prohibit?

The prohibition concerns emotion inference in the named settings.

When did this prohibition begin to apply?

The Commission confirms prohibited-practice rules began in February 2025.

What does the Act’s emotion-recognition definition focus on?

The statutory definition centers on inference from biometric data.

Does the prohibition cover every visible movement or expression?

The Act’s definition has limits that require contextual interpretation.