This analysis describes what Stability AI's agreement states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability and practical outcomes may vary by jurisdiction, enforcement context, and individual circumstances. Read our methodology
The updated terms expand access to audio services for younger users (13+) but require parental consent and supervision for those under 18, shifting responsibility to parents. For developers and researchers, the terms now explicitly restrict using Stability AI outputs to develop, train, or improve competing AI models, with exceptions only when expressly approved in writing. The terms also prohibit removing watermarks or content credentials that identify outputs as AI-generated. For US users, mandatory arbitration now applies; Canada is no longer subject to the arbitration clause. You can review the full updated terms to understand how these restrictions affect your intended use of the services.
View change record →How other platforms handle this
These Terms contain the entire agreement between you and OpenAI regarding the Services and, other than any Service-specific terms, supersedes any prior or contemporaneous agreements between you and OpenAI.
If any provisions from this Policy are in conflict with any applicable laws, such provisions will be replaced by provisions of the same purport that reflects the original intention of the provision, to the extent permissible under the applicable laws.
"You should only use LoRAs that you trust. These LoRAs are user-generated and may be removed by either the trainer or Stability at any time without notice.Excerpt from Stability AI's Terms of Service
Get the weekly coverage letter
Companies change their terms without announcing it. We capture governance documents, compare each new version with the last, and publish what changed, with the source text behind it. One email a week on what we found and what we could not reach.
The clause states: “You should only use LoRAs that you trust. These LoRAs are user-generated and may be removed by either the trainer or Stability at any time without notice.”
ConductAtlas has identified this type of provision across 290 platforms. See the full comparison.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Stability AI.