Smoothwall and AI

[E_G_R2]

Specifically Openai.com but may apply to others but have yet to find an instance.

 

Had three instant safeguarding notification for the likes of self harm and suicide for Openai.com. The link in the alert just goes to a holding page.

 

Question is should we be able to see the question that was typed and the response?

[tom_newton]

Really you would need monitoring for that - the filter will pick up some bits on openai (which I would think carefully about allowing student access to), but almost impossible to get context.

[E_G_R2]

Really you would need monitoring for that - the filter will pick up some bits on openai (which I would think carefully about allowing student access to), but almost impossible to get context.

 

Yes thats what we thought. DNA should do the trick.