It is tempting to move from the observations of the results sections of the various experiments toward normative judgments about what is in fact good or bad about the use of affective computing systems.

However, to do so would be to confuse observing the results of a series of psychology experiments with the formation of notions of what is good concerning affective systems. To simply base guidelines or recommendations upon the results of such empirical investigations would be somewhat like deciding what is good by vote. This is problematic because of the tyranny of the majority; namely that which the largest number says is good is often not the good.

Instead it would be more prudent to make some observations here about the worth of actual observation as compared to professional intuition. As someone who has not only observed but pioneered the development of affective computing systems for many years, it seems that I ought to be able to predict how people respond to them. We have seen that this is certainly not the case and that what has been observed in many cases runs counter to prediction and expectation. This may have implications for methods that ultimately rely exclusively on determinations by the designer, such as Value-Sensitive Design.

Thus, what I can advocate is that professional intuition ought not be relied upon in developing assessments of the ethical implications of affective systems. Instead, situated observation can provide stronger indications of where affective systems are seen as ethical or unethical. Moreover, it is important for these observations to vary social dimensions such as what is at stake and what sort of power relationships exist between individuals.