The Trump administration’s recently signed “One Big Beautiful Bill” allocates $150 billion to the Department of Defense (DoD), with a substantial portion directed toward artificial intelligence (AI) development and deployment initiatives.
The government is aiming to earmark at least $1 billion for offensive cyber operations, according to media reports. In parallel, the administration’s broader AI Action Plan highlights the need for new secure data center infrastructure to support advanced compute capabilities, reinforcing the scale and operational complexity of these AI investments.
While these investments signal a bold leap in defense modernization, they also introduce new risks that must be addressed in tandem.
Chief among them is the growing concern that aggressive AI adoption without corresponding advances in cybersecurity could expose mission systems, data environments and operational infrastructure to critical vulnerabilities.
AI often requires access to vast amounts of sensitive data. And when proper cybersecurity methods are not in place, malicious actors exploit AI to develop sophisticated cyberattacks, scan networks, exploit misconfigurations, steal sensitive data or bypass traditional defenses.
Strengthening cyber defense foundations
Many DoD agencies continue to rely on legacy systems and fragmented security protocols, leaving them unprepared for the demands of integrated, data-driven AI workflows and exposed to escalating cyber threats.
As DoD rapidly integrates AI systems into defense functions, they introduce vulnerabilities that legacy cybersecurity infrastructure is not equipped to handle. These technologies create new attack surfaces including poisoned training data, model extraction and inversion and inference-time manipulation that adversaries can exploit. In many environments, outdated systems and weak controls lead to unsecure workarounds, further increasing risk.
To mitigate risks, AI training and deployment environments must have secure, monitored infrastructure while incorporating cyber hygiene and resilience practices. Resilience practices are proactive cybersecurity methods that assume breaches are inevitable and focus on limiting impact, sustaining operations, and accelerating recovery should an attack occur.
Investing in cyber resilience early in the AI lifecycle not only strengthens mission assurance but also helps reduce long-term technical debt and the likelihood of compromise.
Recommendations for boosting DoD cyber security
AI and cybersecurity must go hand in hand. For every advancement in AI innovation, there must be a cybersecurity-first approach to support and protect it. As the DoD increases its investment in AI, those efforts must deliver measurable returns not only in mission effectiveness but also in security performance.
AI systems should strengthen the department’s ability to defend critical networks and counter cyber-enabled threats from adversaries. To protect these assets, eliminating implicit trust is essential. Cyber hygiene practices must become a non-negotiable requirement for secure operations.
To ensure long-term resilience, program governance must include integrated cyber risk assessments throughout the AI lifecycle, from planning through deployment. The DoD should establish clear cybersecurity performance metrics to track readiness and compliance, supported by dedicated funding for cyber tooling, testing, and skilled personnel aligned with AI efforts.
A recent study by my organization, a cybersecurity company, found that only 42% of organizations have adopted AI specifically to combat ransomware, while more than half (51%) are concerned their organization may experience an AI-generated ransomware attack.
The findings underscore that basic cyber hygiene and resilience practices must remain a priority. Staying secure means making sure the right people have access to the right systems, keeping a close watch on network activity, and quickly identifying any potential threats. It also involves limiting and separating access where necessary, being prepared to contain breaches, regularly updating software and following a mindset of never assuming anything is safe by default. Such practices remain foundational to achieving scalability, enabling efficiency, and ensuring mission readiness.
Recent cybercrimes showcase the dangers. The hackers breached the U.S. National Guard when the cyber group Salt Typhoon “extensively compromised a U.S. state’s Army National Guard network” and “collected its network configuration and its data traffic with its counterparts’ networks in every other U.S. state and at least four U.S. territories” between March and December 2024.
A wake-up call for agencies
The breach serves as a critical wake-up call for agencies to invest in cybersecurity tools and strategies that prioritize cyber hygiene. As agencies accelerate the deployment of advanced AI capabilities, sidelining these foundational cybersecurity efforts would be a serious misstep.
AI represents a strategic advantage for the DoD, but only if it can be trusted, defended and securely maintained throughout its lifecycle.
As the DoD accelerates innovation, it has a critical opportunity to ensure these advancements are not undermined by preventable cyber risks. Without a resilient cybersecurity foundation, even the most sophisticated capabilities risk collapsing under attack. But when officials embed cybersecurity from the start, AI can deliver transformative power to national defense without compromising its integrity.
