dguruwu
commented
2 years ago how would suggested replies be determined? would it be based off a shared data repository of past responses or truly suggestive replies intended to evoke certain unspecified dialogue involving machine learning or neural networks? do we care about privacy (I don't... lol)? could IBM watson assistant be of value? what are our goals in this information exchange?
do you think a UI to suggest replies to help engage in conversations would be useful? having a dialog is probably more persuasive than mass-sending what will appear as propaganda. eventually with enough learning though it could be turned into a "drip feed" campaign where crowdsourced replies are auto-sent.