<p>When an AI system causes harm, to whom do we attribute moral responsibility? The answer might initially appear straightforward—the corporation that constructed the AI is morally responsible. We should treat moral responsibility for harmful AI as we would any other harmful product. Despite its attractiveness, this simple response is too quick. I argue that there are features of the corporate–AI relationship that suggest, in certain cases (with suitably sophisticated AIs), this relationship is more akin to a parent–child relationship than the typical corporate-product relationship with which we are familiar. These features complicate straightforward assignments of moral responsibility. Parental responsibility for juvenile delinquency is not simply responsibility for producing a ‘harmful product.’ Insofar as the features of the parent–child relationship that support this conclusion also arise in certain corporate–AI relationships, a similar conclusion follows with respect to the latter. Corporate responsibility for sophisticated AIs is not simply responsibility for producing a harmful product.</p>