Secrecy vs. Supervision Beyond the Kármán Line: IP Protection, Sovereign Oversight, and the Future of AI-Governed Space Infrastructure

Abstract

This paper tackles a low earth satellite governance paradox beyond the Kármán Line (100 kilometers above sea level): the same proprietary AI that keeps satellites safe also hides the reasoning states need to supervise private actors and assign responsibility. AI black-box compliance is now routine—operators disclose maneuvers but not the internal signals, thresholds, or telemetry transformations—leaving due regard, peaceful-use expectations, and fault analysis to operate on conjecture rather than evidence. The result is an accountability gap across core space-law instruments: Article VI of the Outer Space Treaty presumes continuing supervision; the Liability Convention relies on reconstructable causation; the LTS Guidelines anticipate demonstrable prevention measures. Terrestrial approaches offer partial assistance. The EU’s qualified transparency and the U.S. post-incident auditing travel unevenly off-Earth, and neither framework reliably reaches proprietary on-orbit autonomy. This paper shows with concrete operational scenarios (e.g., dynamic conjunction-thresholding, autonomous servicing approaches), provides an inevitable loss of public-law legitimacy and lack of protection for intellectual property. To address this, the paper proposes a dual-layer disclosure regime that protects legitimate trade secrets while restoring verifiable oversight. Layer 1—Regulatory Safe Rooms: accredited neutral venues conduct confidential code/model/telemetry review under treaty-backed non-disclosure, enabling certification, adversarial stress-testing, and forensic replay without commercial expropriation. Layer 2—Explainability Without Exposure: operators supply functional evidence—validated performance envelopes, adversarial test outcomes, decision bounds—augmented by privacy-preserving attestations (e.g., zero-knowledge proofs) in lieu of source disclosure. Implementation follows a “pressure-valve” path: condition launch licensing, frequency assignments, and mission approvals on participation now; seek UNCOPUOS endorsement later through a model protocol that harmonizes Artemis practices with non-signatories and codifies a TRIPS-compatible IP-Transparency Equilibrium Clause. The payoff is pragmatic rather than utopian: traceability sufficient to make due regard and liability doctrines workable again; incentives preserved for R&D; and a template that can translate to other thin-sovereignty domains (deep-sea, Antarctic, high-altitude autonomy) where algorithmic opacity currently outruns public law.