6.2 KiB
| type | title | author | url | date | domain | secondary_domains | format | status | priority | tags | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| source | Stop Killer Robots / HRW — Alternative Treaty Process Analysis: Ottawa Model and UNGA-Initiated Process as CCW Alternatives | Human Rights Watch, Stop Killer Robots (@StopKillerRobots) | https://www.hrw.org/report/2022/11/10/agenda-action/alternative-processes-negotiating-killer-robots-treaty | 2025-05-21 | ai-alignment |
|
report | unprocessed | medium |
|
Content
Human Rights Watch and Stop Killer Robots have documented alternative treaty pathways outside the CCW framework, relevant given the CCW consensus obstruction by major powers.
Two alternative models:
1. Independent state-led process (Ottawa/Oslo model):
- 1997 Mine Ban Treaty: Independent Ottawa Process led by Canada and NGOs, produced binding treaty banning anti-personnel landmines
- 2008 Convention on Cluster Munitions: Oslo Process, similarly outside UN framework
- Both produced binding treaties WITHOUT requiring major military power participation
- Both succeeded despite US non-participation (US never signed Mine Ban Treaty)
- Mechanism: norm creation + stigmatization + compliance pressure on non-signatories through reputational and market access channels
2. UNGA-initiated process:
- 2017 Treaty on the Prohibition of Nuclear Weapons (TPNW): Initiated via UNGA First Committee
- Adopted by 122 states, in force since 2021
- No nuclear weapons state signed; effectiveness contested
- More inclusive than CCW (doesn't require military powers' consent to negotiate)
Why autonomous weapons are different from landmines/cluster munitions: HRW acknowledges the limits of the Ottawa model for LAWS. Landmines are dumb weapons — the treaty is verifiable through production records, export controls, and mine-clearing operations. Autonomous weapons are AI systems — verification is technically far harder, and capability is dual-use (the same AI that controls an autonomous weapon is used for civilian applications). The technology-specificity of autonomous weapons makes the Mine Ban model harder to replicate.
What's needed for an alternative process to work:
- A critical mass of champion states willing to initiate outside CCW (Brazil, Austria, New Zealand historically supportive)
- Civil society coalition as in previous campaigns (Stop Killer Robots = 270+ NGOs)
- Agreement on scope — prohibit what exactly? Fully autonomous weapons targeting humans without ANY human control? Or also semi-autonomous with insufficient human control?
- A verification architecture (still unsolved technically)
2025-2026 context: May 2025: Officials from 96 countries attended a UNGA meeting specifically on autonomous weapons — the most inclusive discussion to date. The UNGA Resolution A/RES/80/57 (November 2025, 164:6) creates political momentum. Stop Killer Robots advocates that if CCW Review Conference fails in November 2026, the alternative process should begin immediately.
Current status of alternative process: Not formally initiated. Still at advocacy stage. The campaign is explicitly preparing for the November 2026 CCW failure to trigger the alternative process pivot.
Agent Notes
Why this matters: The alternative treaty process is the only governance pathway that doesn't require US/Russia/China consent. But it has two critical limitations: (1) effectiveness without major power participation is limited for a technology those powers control; (2) verification is technically harder than for landmines. The Ottawa model is not directly applicable.
What surprised me: The 270+ NGO coalition (Stop Killer Robots) is larger and better organized than anything in the civilian AI alignment space. The international civil society movement for autonomous weapons governance is more mature than any comparable movement for general AI alignment governance. Yet it has produced no binding instruments after 10+ years. This is evidence that organized civil society alone cannot overcome structural great-power obstruction.
What I expected but didn't find: Any concrete timeline or champion state commitment to initiate the alternative process if CCW fails. The pivot is conditional on CCW failure (November 2026) and still at "advocacy preparation" stage, not formal launch.
KB connections:
- no research group is building alignment through collective intelligence infrastructure despite the field converging on problems that require it — the civil society coalition IS building governance advocacy infrastructure; the gap is in governmental uptake
- AI alignment is a coordination problem not a technical problem — the alternative treaty process is coordination infrastructure for the international layer; it requires the same collective action that domestic governance requires
Extraction hints: "Civil society coordination infrastructure for autonomous weapons governance (270+ NGO coalition, 10-year campaign, UNGA majority support) has failed to produce binding governance because the structural obstacle is great-power veto capacity in multilateral forums, not absence of political will among the broader international community." This would be a specific claim about the limits of civil society coordination as a governance mechanism for great-power-controlled technologies.
Curator Notes (structured handoff for extractor)
PRIMARY CONNECTION: AI alignment is a coordination problem not a technical problem — the alternative treaty process demonstrates that the problem is not technical design of governance instruments but overcoming structural coordination failures among major powers WHY ARCHIVED: Documents the only remaining governance pathway if CCW fails in November 2026. Critical for understanding whether international governance of autonomous weapons AI is a near-term possibility or a decade+ away. EXTRACTION HINT: Compare to the domestic electoral strategy (Anthropic PAC investment): both are attempts to change the political landscape rather than build governance within existing structural constraints. Both face low near-term probability but represent genuine governance alternative pathways.