US Marines Explore Generative AI, Caution Against Blind Trust

As the world races to harness the potential of artificial intelligence (AI), the US Marine Corps is taking a measured approach to integrating generative AI (GenAI) into its operations.

ICYMI: On December 4, 2024, the Marine Corps issued NAVMC 5239.1, a comprehensive policy that outlines how the force plans to adopt GenAI safely and responsibly.

This move highlights the Marine Corps’ commitment to balancing the promise of AI with its inherent risks, ensuring the technology enhances mission effectiveness without compromising operational security.

Balancing Optimism and Caution

The policy seeks to strike a balance between enthusiasm for GenAI and concerns over its limitations.

On the optimistic side, it encourages Marines to explore the potential of GenAI tools.

Commands are explicitly “discouraged from banning the use of GenAI capabilities” and are instead urged to establish governance processes to maximize the benefits while mitigating risks.

However, NAVMC 5239.1 also underscores the importance of vigilance, reminding Marines that GenAI models can “hallucinate”—a term used to describe the generation of inaccurate or entirely fabricated outputs.

To address this, the policy mandates that all outputs be rigorously verified before use.

Key Directives in NAVMC 5239.1

The new policy provides a roadmap for commanders on how to safely integrate GenAI tools into their workflows. Some of the major directives include:

• Accountability and User Identification: Commands are tasked with identifying GenAI developers, system owners, and users to ensure transparency and mitigate risks. This directive stems from concerns raised during the hype cycle of tools like ChatGPT, where some third-party vendors repackaged publicly available AI engines, posing significant security risks. By knowing exactly who is using AI tools and under what conditions, commands can better protect sensitive information.
• Risk Assessment Frameworks: The policy emphasizes the use of established risk assessment frameworks to guide AI implementation. Commands must adhere to best practices outlined in tools like the DoD Responsible AI Toolkit, the National Institute of Standards and Technology’s Risk Management Framework, and the Defense Innovation Unit’s Responsible AI Guidance. This ensures that AI adoption is not haphazard but carefully planned and evaluated.
• Ethical AI Use: The policy requires strict adherence to the Department of Defense’s five ethical principles for AI: responsible, equitable, traceable, reliable, and governable. Commands must track and manage all AI tools in use, ensuring they align with these principles and serve well-defined organizational objectives.

Establishing GenAI Task Forces

One of the most significant aspects of the policy is the establishment of GenAI task forces at major commands. These task forces are designed to take a multidisciplinary approach to evaluating AI tools. Their responsibilities include:

• Assessing existing and emerging GenAI capabilities.
• Determining specific use cases for these tools.
• Creating a list of preferred GenAI solutions that align with mission needs.

By involving subject matter experts in data, knowledge management, AI, and digital operations, these task forces aim to provide commanders with actionable insights. The task forces will also endorse tools that meet operational requirements while maintaining security and ethical standards. Further guidance on task force operations is expected in an upcoming memorandum.

Addressing Security Concerns

Security remains a top priority in the Marine Corps’ approach to GenAI.

The policy’s emphasis on tracking AI tools and their developers reflects a broader concern about safeguarding sensitive data. By mandating the use of trusted frameworks and ensuring that AI tools are traceable, the Marine Corps aims to prevent vulnerabilities that could be exploited by adversaries.

The directive also indirectly addresses lessons learned from the early adoption of GenAI technologies in civilian sectors.

During the rapid rise of tools like ChatGPT, many organizations inadvertently exposed proprietary data by using third-party services without fully understanding their mechanisms.

The Marine Corps’ insistence on accountability and verification is a direct response to such risks.

Encouraging Innovation Within Boundaries

While the policy is cautious, it does not stifle innovation.

On the contrary, it creates a framework for Marines to experiment with GenAI tools in a controlled and secure manner. Commands are encouraged to integrate these tools into their workflows where appropriate, provided they align with organizational goals and maintain operational integrity.

By allowing Marines to explore the potential of GenAI, the policy ensures that the force stays competitive in an era where technology is reshaping the battlefield.

However, this exploration comes with clear boundaries, ensuring that the use of GenAI remains ethical, secure, and effective.

Looking Ahead

The Marine Corps’ cautious yet forward-thinking approach to GenAI reflects a broader trend within the Department of Defense to integrate advanced technologies responsibly.

NAVMC 5239.1 sets the stage for a methodical adoption process that prioritizes security, ethics, and mission alignment.

As the Marine Corps continues to develop its AI capabilities, the upcoming memorandum detailing the operations of GenAI task forces will be a critical next step. By leveraging interdisciplinary expertise and adhering to established guidelines, the Marine Corps is positioning itself to harness the transformative power of GenAI while safeguarding its core values and operational security.

In an age of rapid technological advancement, the Marine Corps’ measured approach serves as a model for how military organizations can innovate without compromising their principles or effectiveness.

You can check out the full NAVMC 5239.1 document here.