
Demanding Transparency in the Government's Software Supply Chain
Published June 6, 2025 • Read time: 5m
Walk into any grocery store in America and you'll find detailed labels on every product. Ingredients, allergens, nutritional facts—all legally required and prominently displayed. Now imagine those same stores filled with unmarked gray boxes, each bearing only the message: "Trust us, it's fine."
You wouldn't accept this for a moment with food. Yet this is exactly how we procure the technology powering our government agencies, military systems, and critical infrastructure. We simply don't know what code lies within the software safeguarding our most vital national interests—a blind spot that has become one of America's most significant security vulnerabilities.
The stakes became painfully clear in December 2021 when the Log4Shell vulnerability emerged—a critical flaw in a single open-source component used by millions of developers worldwide. This glitch triggered a $10 billion global remediation effort and revealed an alarming truth: even our most sophisticated organizations had no reliable way to determine where the vulnerable code lurked in their systems. Today, countless government programs remain unaware of their exposure.
With artificial intelligence, our transparency blindness has become acutely dangerous. Federal agencies have deployed over 700 AI systems with almost no accountability—a Government Accountability Office audit found only five of twenty agencies could identify what's inside their algorithms.
The consequences surfaced starkly last December when researchers discovered a popular AI image generator trained on approximately 1,600 images of child sexual abuse material. This blindness extends to medical devices monitoring patients, defense systems protecting soldiers, and financial networks handling trillions in transactions.
Meanwhile, our adversaries have transformed this transparency gap into a strategic opportunity. Software supply chain attacks have surged 1,300% in just three years, targeting the 80-90% of modern code built from open-source components. Russia, China, Iran, and North Korea have all developed sophisticated capabilities to exploit these blind spots, potentially embedding malicious code in systems controlling everything from fighter jets to missile defense networks—all because we can't adequately inventory what's in our software.
We've seen this story before. In the early 1900s, American food manufacturers routinely adulterated their products with chemical preservatives, toxic dyes, and mysterious fillers—all without disclosure. Consumers ate what they were given, blind to the dangers lurking in their daily bread. The parallels to today's software ecosystem are impossible to ignore: what you don't know absolutely can harm you.
While most food manufacturers fought against transparency, entrepreneurs like H.J. Heinz recognized its transformative potential. As part of a coalition of reform-minded business leaders, Heinz lobbied for the Pure Food and Drug Act of 1906, prominently labeled his ketchup "preservative-free," and opened his immaculate factories to public tours.
This transparency vanguard understood what their secretive competitors didn't: consumer trust creates enduring value. The companies embracing ingredient disclosure thrived in the new regulatory environment, while those clinging to opacity found themselves increasingly irrelevant.
Today's software market needs its own Heinz revolution. Fortunately, a solution is emerging in Software Bills of Materials (SBOMs)—digital ingredient lists that provide machine-readable inventories of every component within an application. Like nutrition labels, SBOMs promise transparency and accountability throughout the digital supply chain, enabling users to finally see what's inside the code they depend on.
The federal government has recognized this parallel to history. In May 2021, President Biden issued Executive Order 14028, which laid groundwork for requiring SBOMs across federal agencies—our digital equivalent to the 1906 Pure Food and Drug Act. The Army, Navy, and Air Force have implemented their own requirements, demanding suppliers provide these software "nutrition labels" with their products. While these are promising steps, adoption remains spotty and implementation inconsistent across government.
Even more alarming is our headlong rush into artificial intelligence without equivalent transparency safeguards. While traditional software remains dangerously opaque, AI systems introduce entirely new dimensions of risk, requiring specialized AI-SBOMs to document their models, training data, and unique components. Without this visibility, we're deploying black boxes to make consequential decisions throughout government—boxes whose contents remain a mystery even to their operators.
Predictably, industry critics insist that such transparency is too complicated, too expensive, or would reveal proprietary secrets. It's an echo of history: the exact arguments food manufacturers deployed against ingredient labels a century ago. Yet history has rendered its verdict: the market ultimately rewarded companies that embraced transparency while punishing those that resisted it. In technology, as in food, consumers will demand to know what they're consuming.
Some pioneers are already leading the way. Lockheed Martin has emerged as a trailblazer in SBOM adoption throughout its vast software portfolio, recognizing that transparency builds both security and customer trust. Yet they remain a striking exception in the defense industrial base, where most contractors continue fighting against the very transparency their government customers increasingly demand. Their resistance echoes food manufacturers who once fought against disclosing artificial colors and chemical preservatives.
Legacy systems pose real challenges—how do you inventory decades-old code written by long-retired developers? Yet these legitimate difficulties with yesterday's software cannot justify blindness to tomorrow's. The wisdom about planting trees applies perfectly: the best time to demand software transparency was years ago. The second-best time is now.
For government agencies, the path forward is clear: reject the black box era. Make transparency non-negotiable in every contract. Require AI-SBOMs for all machine learning systems. Set a standard where software controlling nuclear infrastructure faces at least the same disclosure requirements as breakfast cereal.
Our adversaries already target the very transparency gaps we permit, turning our digital blind spots into their entry points. The nation that masters software transparency first won't just build more secure systems—it will define the next era of technological leadership. The transparency revolution is coming to software. America can lead it or be led by it.
Evan Loomis is a senior advisor to DARPA, co-founder of ICON, and investor at Overmatch Ventures, where he focuses on critical technology, space, and defense. Marc Frankel is the CEO and co-founder of Manifest Cyber.

Why Iran Cannot Simply Surrender
By Eric Garland • RealClearDefense
The Formation Left Us Behind
By Mark W. Castillon • RealClearDefense
Can Commanders Tell When Artificial Intelligence Is Wrong?
By Jonathan K. Corrado • RealClearDefense
The Summit as Mirror
By Stephen Blank • RealClearDefense
The False Choice That Could Get Us Killed
By Adam Strickland • RealClearDefense

Why Iran Cannot Simply Surrender
By Eric Garland • RealClearDefense
U.S. Navy Cuts Submarine Maintenance Delays by 40%
By Ethan Gossrow, Naval News
Air Force Bets on Hybrid-Electric Aircraft for Transport
By L.C. Williams, Defense One
Roosevelt Joins 2 Other Carriers in Mideast As Pressure on Iran Rises
By Stars & Stripes
Fight Now? Time to beef up III MEF
By Marine Corps Compass Points
