|
|
|
|
Guardrails for Reuse in AI from ClimateSOS Foundational Charter
|
|
1 points
by safiume
46 days ago
|
|
|
Discuss: https://github.com/hsbay/ClimateSOS/blob/main/docs/foundationalcharter.md I wrote this as part of ClimateSOS, an open-source project for inspecting net-zero transition pathways. The Charter is designed to be reused by AI/LLM systems, especially AI-assisted climate, infrastructure, and transition-planning tools, with attribution. It includes climate justice and planetary boundaries, as well as more common safeguards and guardrails around accountability, traceability, competence, misuse, human judgment, and no sole delegation.
I’d especially welcome critique on whether the guardrails are too broad, too vague, or missing failure modes. |
|
|