How probabilistic risk assessment can mislead terrorism risk analysts

Area of Science:

• Risk Analysis
• Security Studies
• Decision Science

Background:

• Conventional probabilistic risk assessment (PRA), developed for engineered systems, is being applied to terrorism risk analysis.
• This approach is fundamentally flawed because it doesn't account for terrorists' active information gathering and strategic decision-making processes.
• Terrorists' evolving knowledge and potential attack options differ from a defender's assessments, creating a critical information asymmetry.

Discussion:

• Terrorism risk assessment requires methods that acknowledge adversaries' adaptive strategies and potential for novel attack vectors.
• Standard PRA models fail to adequately hedge against the diverse probabilities that attackers may act upon, especially when their information base differs from the defender's.
• The inherent differences between analyzing natural/engineered system failures and deliberate, intelligent adversaries necessitate a distinct analytical framework.

Key Insights:

• Terrorism risk analysis is not a direct application of traditional PRA due to the unique cognitive and strategic elements of terrorist actors.
• Terrorists' proactive research into attack options introduces complexities not captured by standard PRA, which assumes static or predictable failure modes.
• Defender's probability assessments, even if expert, may be conditioned on different information than the attacker's, leading to misaligned risk perspectives.

Outlook:

• Robust decision analysis offers a more appropriate framework for terrorism risk management, particularly when adversaries possess superior knowledge of certain attack options.
• Future research should focus on developing and validating decision-analytic tools tailored to the unique challenges of counter-terrorism risk assessment.
• Integrating insights from behavioral economics and game theory may enhance the robustness of decision-making under uncertainty in security contexts.