AI in Security: Expert insights on leveraging technology, governance and global competition
As artificial intelligence (AI) becomes more integrated into security operations, its potential to reshape threat detection, risk management and response strategies is growing at an unprecedented pace.
Yet, as AI evolves, organizations must navigate governance challenges, regulatory landscapes, and the ongoing race for technological dominance. The global competition for AI supremacy is often compared to the Space Race, with countries and corporations striving to develop the most powerful and efficient AI models.
One of the latest U.S. initiatives to accelerate AI innovation is Stargate, a multi-billion-dollar effort aimed at ensuring the nation remains at the forefront of AI research and development. Stargate is focused on addressing key infrastructure challenges, including chip manufacturing, data center scalability, and cloud computing capacity — all of which are critical to supporting the AI boom. By streamlining AI advancements and fostering collaboration among leading experts, Stargate seeks to position the U.S. ahead of competitors like China, who are also investing heavily in AI dominance.
For the electronic security industry, these advancements hold significant implications. As AI-powered security solutions — such as video analytics, threat detection and automated incident response — become more prevalent, integrators and end users must consider how evolving AI infrastructure and regulations will shape their adoption strategies. Enhanced AI capabilities could reduce false alarms, improve response times and enable predictive threat modeling, but without proper governance and security measures, organizations risk exposing sensitive data or deploying unvetted technologies.
To explore these critical issues, SecurityInfoWatch spoke with Sean Tufts, managing partner for critical infrastructure and operational technology at Optiv, a provider of cybersecurity programs from strategy and managed security services to risk, integration and technology solutions. Tufts shares insights on how systems integrators can harness AI for security applications, the governance considerations organizations must address, and how initiatives like Stargate are shaping the United State’s position in the global AI race.
Maximizing AI’s Potential in Security Integration
SIW: How can systems integrators leverage advancements in AI to enhance the security and efficiency of critical infrastructure?
Tufts: The “signal to noise” ratio that our defenders must wade through is a true challenge. The average enterprise has 70 different security tools. All of them create a log, alert, alarm, or report. The last five years has seen progress on harmonizing and integrating these tools (mostly through SOAR platforms). The next level will leverage AI to improve our SOC analyst’s data and the ability to get them much faster.
SIW: What are the key considerations for systems integrators when incorporating AI technologies into existing security frameworks?
Tufts: Governance is the key to AI. Opening up an org to a LLM without guardrails is a security disaster. Narrowing the AI use cases and limiting the related data ingress and egress is key. Users can get frustrated by the feature-lite uses, but that also focuses their efforts on the outcomes the business is looking to gain.
For example, using predictive AI tools for fraud detection is a really simple and helpful use case. But we need to limit the financial information fed to the tool to only things that are associated with identify fraud; things like date, times, names, account numbers, etc. We need to avoid putting the full financial stack into the model. The fraud platform does not need unreleased earnings figures.
SIW: How can systems integrators stay ahead in the AI race and ensure they are providing cutting-edge solutions to their clients?
Focus, focus, focus. AI use cases have short half-life. Meaning, engineers want to prove out cool new projects. They often move on to the next one before a full evaluation and documentation has been completed. AI has been very vulnerable to “shiny object syndrome.” Having a firm vetting to determine which features are “hobbies” verse a core project will help integrators bring new breakthroughs to market with limited wasted time.
Navigating AI Adoption and Compliance Challenges
SIW: What steps should end-user organizations take to assess the potential benefits and risks of adopting AI technologies in their operations?
Tufts: Focusing an enterprise on use case-driven AI has yielded the best outcomes. Successful programs have a use case development flow that includes a formal intake, capacity and priority planning, model development, multi-stakeholder review and change management based on roll out, metric review and decommissioning planning. This keeps security at the front of the conversation and embedded throughout the lifecycle.
SIW: How can end-user organizations ensure they are compliant with evolving AI regulations while still fostering innovation?
Tufts: Limiting the number of AI platforms in use is the easiest way to help. Having data in multiple third-party AI platforms will cause many governance issues, especially for global rollouts.
SIW: What role do investments in initiatives like Stargate play in positioning the U.S. as a leader in the AI race, and how can end-user organizations benefit from these advancements?
Our infrastructure was not prepared to handle these AI data workloads. Our underlying infrastructure of chip manufacturing, data center scale, available fiber data highways, cloud service providers, power supply and capacity, and cooling systems need to re-work their roadmaps to handle this current surge.
Stargate was made to collaborate on bold — and fast — solutions. The underlying messaging behind Stargate seems to be: “We want to build as fast as China. Let’s get the best people in the world and get out of their way.”
