Recently started using AI so I am likely to be missing a setting.
With the upcoming retirement of Google Perspective API moved forward with the much avoided AI implementation.
Both Spam and Toxicity are working but the same post is being flagged as both spam and toxic. Is this normal and expected behavior?
Nomenclature feedback: In the AI applied flag context “Flag rejected” is confusing. The post is rejected and flagged as rejected or was the initial AI flag rejected?
Adding another example of using both Spam detector persona (preconfigured) and a Toxicity persona.
The post triggers as both spam and toxic resulting in two flagged entries with very different response options and outcomes.
Question: Given the likelihood of most any non-private forum needing both spam and toxic post prevention would it make sense to combine Spam and Toxicity personas into one preconfigured persona?
One post two flags
Is this post spam? Yes
Is this post spam? No
Approve this post? Yes with no options or No
Looks like Sam had a similar suggestion that I missed earlier.
After spending more time working with this and giving it some thought it is possible the double flagging, spam and toxic, can be lessened if I can make some appropriate adjustments to the Toxicity Triage Persona. In my experience toxicity is more likely to occur in a non-new member as opposed to a spammer.
With that said I think a look at
is warranted.
Also, combining Spam and Toxicity, or including Toxicity similar to Spam module seems like a reasonable feature request and discussion so moving to Topics tagged ai
