Precautionary Principle
A risk management approach requiring protective action against potential harms even when scientific certainty is lacking.
Definition
Precautionary Principle is a risk management approach holding that when an activity raises threats of harm, precautionary measures should be taken even if some cause-and-effect relationships are not fully established scientifically. In AI governance, it supports proactive regulation of potentially harmful technologies.
EU Application: The precautionary principle is enshrined in EU law through Article 191 TFEU (Treaty on the Functioning of the European Union). The EU AI Act's approach to prohibiting certain AI practices (Article 5) and requiring pre-market assessment of high-risk systems reflects precautionary thinking—regulating potential harms before they fully materialize.
Key Elements:
- Threat of harm: Reasonable grounds to believe harm may occur
- Scientific uncertainty: Full causal evidence may not exist
- Precautionary action: Protective measures taken nonetheless
- Proportionality: Measures appropriate to potential risk
- Burden of proof: May shift to those proposing the activity
Jurisdictional Variations:
- EU: Strong embrace of precautionary principle; embedded in AI Act's prohibitions and pre-market requirements
- US: Generally more risk-tolerant approach; innovation-focused policies with post-market correction
- International: Rio Declaration Principle 15 articulates precautionary approach for environmental protection
Application to AI: The precautionary principle justifies prohibiting certain AI applications (social scoring, subliminal manipulation) without waiting for proof of harm, and requiring safety assessments before deployment of high-risk systems.
Related concepts: Risk Management System, Prohibited AI Practices, Conformity Assessment
Sources
- •EU Treaty Article 191
- •Risk Regulation Principles