Software Engineering Institute (SEI) Podcast Series · Members of Technical Staff at the Software Engineering Institute

Delivering Next-Generation AI Capabilities

·30 min·1 clip

We still don't understand how modern AI systems actually work.

Bukovic opens with the SEI's institutional frame and introduces Gaston as the director of the Artificial Intelligence Division. From there, the discussion turns to the question sitting underneath the whole episode: how defense and national security organizations can adopt AI capabilities that are not only impressive, but also scalable, reliable, and assured. Gaston brings long experience at the intersection of national security and technology. He explains that the SEI's value comes from standing between the commercial world and government users, where it can show what these technologies can do, show the right ways to use them, and help translate them into larger systems that still have to live alongside legacy capability or new capability under development. The conversation keeps coming back to the gap between what is possible and what is safe. Gaston notes that AI systems can be powerful while still being complex, not fully understood, and easy to use in unintended ways, so the engineering problem is to set them up so they can be trusted in serious operational settings. That tension is treated as the main issue, not a side note. Bukovic repeatedly brings the discussion back to practical adoption, asking how the SEI can help players of all sizes in the defense industrial base with acquisition and deployment of local systems without losing sight of trust and mission needs. Gaston answers by emphasizing expertise. He describes the value of having people who are expert first users of these technologies, because that is what lets the SEI demonstrate useful patterns instead of abstract enthusiasm. The episode also frames AI as unavoidable. Once it becomes part of the lifeblood of defense missions, the question is no longer whether to engage with it, but how to do so in a way that is dependable, integrated, and fit for purpose. That leads into the national conversation about trustworthy systems. Gaston points to efforts aimed at building ecosystems of valuation that can support reliability, explainability, and security, and he places the SEI's work on testing and assurance inside that broader effort. The focus stays on systems, not slogans. Rather than treating AI as a standalone tool, the conversation keeps folding it back into acquisition, deployment, interoperability, and the realities of defense environments where old and new capabilities have to work together. The tone stays measured throughout. Bukovic's questions leave room for detail, but they also keep pulling the answers into plainer language so the core logic stays visible: AI can be useful, but it has to be engineered for trust, not assumed into existence. The result is a pragmatic conversation. Listeners come away with a clear sense of why the SEI sees its role not as hyping AI, but as helping the defense community adopt it with the discipline required for mission-critical use.

As heard by us

A sober look at how defense AI earns trust and moves into real deployments.

This episode keeps its focus on the hard part of defense AI: getting capability into real missions without hand waving away the work of assurance, reliability, and testing. It treats those issues as the main event, which keeps the discussion grounded.

Read the full review in PlayNext →

Why you'd press play

Want a practical briefing on bringing trustworthy AI into defense work?

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Software Engineering Institute (SEI) Podcast Series