Florian Richter

Computational or machine ethics examines how morality could be implemented in machines and represented computationally. In the cognitive sciences, different architectures (rule-based and connectionist) have been proposed. Nevertheless, such approaches have limitations when incorporating ethical claims. With the rise of artificial intelligence, new ways of simulating moral expertise seem possible, where chatbots can serve as dialogue partners for moral orientation. However, as I will argue, such systems lack a theoretical grounding in metaethics, as do the former approaches. In order to align such systems ethically and ground them conceptually, a topography of moral competence must be developed. By outlining this topography, I will show some directions for further research and how such systems might enhance us morally.

Invasive brain-computer interfaces (BCIs) pose a challenge for describing human-system interactions because they involve a depth of engagement that is difficult to disentangle. This is particularly problematic when clear authorship of decisions is needed. This presupposes that such devices might also decide autonomously in certain situations. Although the use and implementation of such devices are highly disputed in practical ethics, they seem to offer at least therapeutic benefits for persons with, e.g., Parkinson’s disease. That is why it is indispensable to develop a conceptual framework that allows a description of the system’s interactions with the human agent. The effects of brain-computer interfaces on patients are assessed empirically in studies and hypothesized on a more speculative level via thought experiments. A model to operationalize dimensions of agency, such as responsibility and privacy, has been proposed. Nevertheless, this model is inchoate if it is not related to autonomy, because it is widely assumed that autonomy is not only a characteristic of the human agent but also of the technological system (at least to a certain degree). Consequently, a more complete model is needed that classifies autonomy at different levels (operational, strategic, and moral). Technological systems can only operate on the first two levels because their behavior is based on dispositions within the scope of alethic modality, whilst human agency is based on deontic structures that are expressed via normative attitudes and statuses that cannot be reduced to alethic modality. This crucial difference can be used to develop methods for evaluating the ethical side of BCIs.

Invasive brain-computer interfaces (BCIs) pose a challenge for describing human-system interactions because they involve a depth of engagement that is difficult to disentangle. This is particularly problematic when clear authorship of decisions is needed. This presupposes that such devices might also decide autonomously in certain situations. Although the use and implementation of such devices are highly disputed in practical ethics, they seem to offer at least therapeutic benefits for persons with, e.g., Parkinson’s disease. That is why it is indispensable to develop a conceptual framework that allows a description of the system’s interactions with the human agent. The effects of brain-computer interfaces on patients are assessed empirically in studies and hypothesized on a more speculative level via thought experiments. A model to operationalize dimensions of agency, such as responsibility and privacy, has been proposed. Nevertheless, this model is inchoate if it is not related to autonomy, because it is widely assumed that autonomy is not only a characteristic of the human agent but also of the technological system (at least to a certain degree). Consequently, a more complete model is needed that classifies autonomy at different levels (operational, strategic, and moral). Technological systems can only operate on the first two levels because their behavior is based on dispositions within the scope of alethic modality, whilst human agency is based on deontic structures that are expressed via normative attitudes and statuses that cannot be reduced to alethic modality. This crucial difference can be used to develop methods for evaluating the ethical side of BCIs.

Technological implementations in the educational context seem to contribute to competence acquisition through personalized learning. For instance, intelligent systems are embedded in the learning environment in smart classrooms. Implementing these systems is widely accepted, and different technical solutions have been proposed, but the educational context still needs to be thoroughly scrutinized. The examined approaches fail to capture the complexity of the learning experience and underdetermine the pedagogical aspect. Therefore, directions are indicated and discussed to clarify the educational context of technological implementations and show how intelligent systems can support personalized learning in an ethically acceptable way.

The development of autonomous weapons systems (AWSs), which would make decisions on the battlefield without direct input from humans, has the potential to dramatically change the nature of war. Due to the revolutionary potential of these technologies, it is essential to discuss their moral implications. While the academic literature often highlights their morally problematic nature, with some proposing outright banning them, this paper highlights an important benefit of AWSs: protecting the lives, as well as the mental and physical health of soldiers. If militaries can avoid sending humans into dangerous situations or relieve drone operators from tasks that lead to lifelong trauma, this obviously appears morally desirable – especially in a world where many soldiers are still drafted against their will. Nonetheless, there are many arguments against AWSs. However, we show that although AWSs are potentially dangerous, criticisms apply equally to human soldiers and weapons steered by them. The combination of both claims makes a strong case against a ban on AWSs where it is possible. Instead, researchers should focus on mitigating their drawbacks and refining their benefits.

Smart technologies should make cities, governments, and societies not only more efficient and sustainable but also more livable. First, the conceptual framework of smart cities and societies is reviewed. At first glance, it seems promising to implement AI-based systems to address the latent needs of citizens. However, which conception of justice this presupposes has not been sufficiently scrutinized. Second, different conceptions of justice are shortly discussed. It is argued that conceptions of justice are context-dependent, i.e., while in one context, one conception of justice could be considered ethically acceptable, in another context, it might be considered unacceptable. Third, the limits and ethical implications of operationalizing justice via algorithms are analyzed.

Rule-following is a normative doing and therefore needs to be reconsidered in a metaethical framework. Rule-following will be discussed in the light of cognitivism and non-cognitivism. It will be shown that neither cognitivism nor non-cognitivism are sufficiently good accounts for conceptualizing rule-following, because they are held captive by a quasi-mechanistical picture of rule-following. This idea stems from Stanley Cavell´s and John McDowell´s approach to rule-following. McDowell appeals to the idea that we participate in “shared forms of life” and therefore are able to follow rules correctly. I believe that his quietist view can be amplified by conceptualizing different modes of following a rule. We can also follow a rule e.g. ironically and still grasp the conceptual content of the rule adequately.