Bernhard Koch first points out the capabilities and also the limits of the capabilities of AI ethics. Some terms that are currently common in AI ethics are far less normatively effective than they appear in public rhetoric. Above all, the specificity of the moral problem is overlooked if one only pays attention to the effects in physical reality. It is more exciting to look at the effects in our moral practice itself. Here, in particular, generative AI shows what Sven Nyholm calls a “credit-blame asymmetry”, which – if one thinks about it further psychologically and ethically – can lead to a decline in a culture of moral praise and a change in the culture of moral blame. This also applies to military medicine. However, the dual role in which military physicians find themselves may have a kind of refraction that continues to require independent reflection and judgment and thus also enables praise and blame.