Problem being solved
StopNCII asks participating platforms to report back on hash matches, whether matched content was confirmed as NCII and actioned, or was a false positive. This feedback loop is what keeps the StopNCII hash list accurate and is part of being a good participant in the program.
Today, when content matches a StopNCII hash (via HMA) and a moderator makes a decision in Coop, that outcome stays in Coop. Reporting back to StopNCII is a separate manual step, which in practice means it often doesn't happen.
Proposed solution
When a moderator decision is made on a job whose signal originated from a StopNCII hash match, Coop (or HMA, if that's the better layer for it) automatically submits the corresponding feedback to StopNCII. e.g. confirmed/actioned vs. false positive, mapped from the decision taken.
Ideally this would be configurable per integration (opt-in, with the platform's StopNCII credentials), similar in spirit to the existing NCMEC reporting flow.
Alternatives considered
- Manual feedback via StopNCII directly: works, but relies on moderators doing double data entry, so coverage is poor.
- A periodic export from Coop that platforms transform and submit themselves: still manual glue code for every platform.
Anything else
This came up from a platform using Coop + HMA with the StopNCII hash list who want to participate in the feedback loop without building custom tooling around Coop.
Screenshots
No response
Checklist
Problem being solved
StopNCII asks participating platforms to report back on hash matches, whether matched content was confirmed as NCII and actioned, or was a false positive. This feedback loop is what keeps the StopNCII hash list accurate and is part of being a good participant in the program.
Today, when content matches a StopNCII hash (via HMA) and a moderator makes a decision in Coop, that outcome stays in Coop. Reporting back to StopNCII is a separate manual step, which in practice means it often doesn't happen.
Proposed solution
When a moderator decision is made on a job whose signal originated from a StopNCII hash match, Coop (or HMA, if that's the better layer for it) automatically submits the corresponding feedback to StopNCII. e.g. confirmed/actioned vs. false positive, mapped from the decision taken.
Ideally this would be configurable per integration (opt-in, with the platform's StopNCII credentials), similar in spirit to the existing NCMEC reporting flow.
Alternatives considered
Anything else
This came up from a platform using Coop + HMA with the StopNCII hash list who want to participate in the feedback loop without building custom tooling around Coop.
Screenshots
No response
Checklist