Imagine this: a satellite’s onboard AI misfires, rerouting its orbit — not due to a malfunction, but because its software supply chain was quietly compromised months earlier. Or an adversarial algorithm manipulates imagery from space-based sensors, eroding trust in what we see from above.

These are no longer theoretical concerns — they’re exactly the kinds of scenarios experts gathered to discuss at Space Cybersecurity With AI (SCYWAI).

On March 18–19, 2025, USC hosted this two-day event bringing together researchers, defense officials, industry leaders and students to explore one of the most urgent and complex challenges of our time: securing space systems in the age of generative AI. Sponsored by the National Science Foundation, the first day was held at USC’s main campus and day two took place at USC Viterbi’s Information Sciences Institute (ISI). 

Participants from organizations including NASA’s Jet Propulsion Lab, Boeing Corporation, Lockheed Martin, Northrop Grumman Corporation, Cisco Corporation, The Aerospace Corporation, U.S. Space Force SpaceWERX, Space ISAC, SixGen, Cyware, Chainguard, Duality Technologies, Mandala Space Ventures and AIN Ventures reflected the cross-sector collaboration at the heart of the event. Sessions covered topics like anomaly detection, autonomous response systems, regulatory hurdles and how to build secure AI pipelines for space systems. Attendees spanned disciplines, sectors and agencies — and one clear theme rang true: this challenge can’t be tackled alone.

Why USC/ISI Is Built for This Moment

CREST — the Center for Research in Space Technologies — is a university-wide initiative; a collaboration among the Viterbi School of Engineering (including ISI); the Dornsife College of Letters, Arts, and Sciences; and the Sol Price School of Public Policy. CREST was established to advance the understanding, exploration, and use of space. Its mission: to translate emerging risks into actionable strategies — and bold ideas into real-world infrastructure.

“It is critical that we engage innovators and thought leaders in the space domain,” said Andrea Belz, USC Viterbi vice dean for transformative initiatives and director of CREST. “The new space economy will only grow in its importance to the economy and our daily lives. L.A. is fortunate to have so much talent in this field.”

Alefiya Hussain, a Principal Research Scientist at ISI and deputy director of USC CREST, who led the planning of the workshop, noted that ISI is “positioned to do both open research as well as sensitive classified research” — an important combination that enables collaboration across academic, industry, and government sectors. With strong federal partnerships and close ties to Southern California’s thriving aerospace sector, USC is well positioned equipped to lead these conversations.

“We’re in a moment where everything is changing,” said Hussain. “Launch costs have come down so dramatically that space is no longer just for governments and defense contractors. That opens up huge possibilities — and huge vulnerabilities.” At the same time, the rise of generative AI has introduced powerful new tools and new risks. “It’s a very complex domain,” Hussain said. “You’re trying to apply techniques that need to work in resource constrained environments.”

Keynotes: “The future is something we can both predict and create”

Charles Brown, deputy director, U.S. Space Force SSC S6, delivered the first keynote, highlighting the double-edged nature of artificial intelligence. “The same technologies that empower us also present unprecedented security risks,” he said. Brown outlining the Space Force’s five core competencies: space security, combat power projection, space mobility and logistics, information mobility and space domain awareness. “Timely delivery of those space capabilities is even more critical to the nation.”

He described the Department of the Air Force’s investment in AI tools that “augment human decision making, enabling faster and more informed decisions,” and its efforts to build a workforce capable of rapidly developing and deploying secure, reliable applications. “In the world of AI and cybersecurity,” he said, “the future is something we can both predict and create together.”

Charles Brown / Photo credit: USC

In the second keynote, delivered by Neil Siegel, IBM Professor of Engineering Management at USC and former Chief Technology Officer at Northrop Grumman, Siegel offered a candid reflection on AI’s readiness for high-stakes missions. Drawing on decades in national security, he cautioned that today’s AI systems often hide uncertainty, lack clear boundaries, and can’t explain their decisions — all critical gaps when lives or policy decisions are on the line. “New technologies can have advantages,” he noted, “but they also almost always have unplanned adverse behavior.”

Neil Siegel / Photo credit: USC

Inside the Room: A Student’s Perspective

For Gabriel Rodriguez, a computer science master’s student at Cal Poly San Luis Obispo, the workshop offered a chance to see his research on securing software supply chains for CubeSats in a real-world context. CubeSats — small, low-cost satellites used for research and education — are increasingly important targets for cybersecurity, as their rapid development cycles and open architectures make them especially vulnerable.

“The sessions were inspiring,” Rodriguez said. “It was validating to realize how many people are working on similar problems, and how my work might actually make a difference. I especially appreciated the openness of the discussions — it wasn’t just theory, it was about what’s happening now and what needs to happen next.”

Looking Ahead

The workshop made clear that cybersecurity for space demands new kinds of thinking. Generative AI, while powerful, isn’t a plug-and-play solution — especially in high-risk, high-stakes environments like space systems. Addressing these challenges will require new models of trust, verification and collaboration.

Organizers described the event as the beginning of a longer-term effort to draft a national roadmap for space cybersecurity in the AI era. There was wide agreement that the U.S. must move faster — but also a shared optimism that with the right people at the table, and the right questions being asked, solutions are within reach.

The conversation reflected a larger shift: space is no longer just about who can get there first — it’s about who can keep it secure, and that effort will take everyone.

Published on March 26th, 2025

Last updated on March 28th, 2025