How AI Might Affect Decisionmaking in a National Security Crisis

Overview:  

This article was written by Christopher S. Chivvis and Jennifer Kavanagh at Carnegie Endowment for International Peace.

​

• AI systems can both accelerate and complicate decision-making in national security scenarios.
• Overconfidence in AI recommendations could lead to groupthink and potentially dangerous misperceptions.

​

Key Quotes:  

• "AI-enabled systems can help accelerate the speed of commanders’ decisions and improve the quality and accuracy of those decisions."
• "In reality, AI systems are only as good as the data they are trained on, and even the best AI have biases, make errors, and malfunction in unexpected ways."

​

What They Discuss:

• The proliferation of AI in national security could slow decision-making because AI systems produce additional data that need to be evaluated.
• AI’s potential to create uncertainty in crisis situations involves deepfake videos and potentially misleading information.
• AI might challenge existing groupthink in decision-making settings by offering out-of-the-box ideas but could also entrench it if decision-makers over-rely on AI recommendations.
• The development of AI tools by well-funded agencies could disturb the balance of influence among key governmental bodies like the Department of Defense and Intelligence Community.
• Misjudging adversary actions influenced by AI systems could escalate crises due to the risk of miscalculation.

​

What They Recommend:

• Implement thorough training for policymakers on AI systems to understand their limits and capabilities.
• Establish an AI governance regime similar to arms control to manage and reduce risks of AI deployment in military contexts.
• Foster international cooperation, especially between the U.S. and China, on AI safety and governance measures.

​

Key Takeaways:

• AI has the dual potential to both streamline and complicate crisis decision-making processes.
• Training and prior experience with AI tools are crucial for their effective and safe use.
• Establishing clear norms and agreements on AI use is important for reducing the risk of misperceptions and unintended escalations.
• Policymakers must be wary of AI’s potential to sway groupthink and maintain a balanced approach incorporating human judgement.

​

This is a brief overview of the article by Christopher S. Chivvis and Jennifer Kavanagh at Carnegie Endowment for International Peace. For complete insights, we recommend reading the full article.