National Security, AI & Ethics

Introduction

The U.S. Department of War has recently entered into a partnership with AI firms such as OpenAI to deploy advanced models within classified networks. The agreement was finalised after Anthropic declined to participate, citing concerns over compromising its core safety principles. Specifically, Anthropic emphasised its firm stance against domestic mass surveillance and its commitment to ensuring meaningful human accountability in the use of force, including in the context of autonomous weapon systems.

A similar situation may emerge in India, where rapid AI integration in defence, supported by initiatives like the Defence AI Council and the Defence AI Project Agency, and which may be driven by private firms.

Ethical values upheld by Anthropic

• Trustworthiness: By refusing to sign deal, the company showed its commitment to protecting individuals' data from potential misuse, thereby reinforcing public confidence.
• Courage: The willingness to take a principled stand against government pressure despite potential financial and strategic costs.

• Non-maleficence: It has avoided harm by refusing partnership in applications that could enable misuse or rights violations.

Key Ethical Dilemmas in this matter

• National Security vs. Moral Responsibility: Should firms prioritise national interest or their own ethical commitments?
  • In the Anthropic case, Integrity is reflected in a company's principled refusal to compromise its ethical commitments, even at the cost of losing lucrative government partnerships.
• Utilitarianism vs. Rights-Based Ethics: As per utilitarianism approach (Jeremy Bentham, John Stuart Mill), the use of AI in defence by prioritising the greatest security benefit for the majority can be justified. 
  • However, rights-based ethics cautions against such actions as they can potentially be misused. 
• Duty Ethics vs. Consequence Ethics: Can compromising individuals' data and privacy ever be justified in the name of national security and the greater good?
  • The U.S. government and OpenAI adopted a consequentialist approach, prioritising national security outcomes, whereas Anthropic followed a deontological stance, refusing to cross certain ethical boundaries despite potential monetary benefits.
• Corporate Autonomy vs. State Control: Do Private firms have the right to decide how their technologies are used or they have to abide by the government?

• Autonomous Systems vs. Accountability: If an AI system malfunctions, commits an error, or violates international humanitarian law, attributing legal and moral responsibility becomes extremely difficult, allowing commanders to potentially deflect blame onto the machine.
  • Many advanced AI systems lack transparency and explainability, operating as black boxes where users cannot fully understand how the system reached a specific conclusion. 
• Surveillance Tools vs. Liberty: Deploying advanced AI within classified networks risks enabling large-scale surveillance based on extensive data about individuals' behaviour, preferences, and personal traits.

Conclusion 

The case highlights the urgent need to balance technological advancement with ethical responsibility, ensuring that the pursuit of national security does not undermine fundamental rights and human dignity. As AI becomes central to warfare, robust frameworks rooted in accountability, transparency, and human oversight are essential.