Even if I could develop a morality engine and install it in every device, system, and process I've ever worked on, I don't think I would. Not only is it too comnplex a problem, it subverts the morals of the user and substitutes my own. And I Know I don't have the far ranging vision to appreciate the fine points of every potential future situation to evaluate them properly. It is hard enough to do that well in real time, with all or most of the facts and evidence present for examination.
Any engineer, actiing responsibly, can take or refuse a job based on the knowledge at hand, and whatever moral framework may seem to apply. But predicting the future uses as well, no. It has been generally ruled out and rightly so. To do otherwise assumes people of the future are incapable of seeing their own situation and evaluating it for themselves. That kind of deprecation is as bad or worse than the kind of ancestor worship that says our forebears were smarter, wiser, more moral, etc than we are today. Still wrong, but at least in looking back we have evidence to back up (some of) the claims.