Etye Steinberg

For each person, some actions are unthinkable: performing them requires crossing a line that one's conscience cannot allow crossing. This article explores what such unthinkability is. In doing so, it introduces a novel categorization of theories of action and practical reason. The article argues that an action is unthinkable if and only if the agent judges that she should never treat any consideration as a reason in favor of performing this action. This view meets two important tests. The first is ‘Failure’: a theory of unthinkable action must allow for the conceptual possibility of doing the unthinkable and explain what is special about the failure implicated in doing the unthinkable, especially in comparison to the failure implicated in akratic action. The second is ‘Incapacity’: a theory of unthinkability must distinguish between not doing something because it is unthinkable and not doing something because of an extremely negative psychological reaction.

For each person, some actions are unthinkable: performing them requires crossing a line that one's conscience cannot allow crossing. This article explores what such unthinkability is. In doing so, it introduces a novel categorization of theories of action and practical reason. The article argues that an action is unthinkable if and only if the agent judges that she should never treat any consideration as a reason in favor of performing this action. This view meets two important tests. The first is ‘Failure’: a theory of unthinkable action must allow for the conceptual possibility of doing the unthinkable and explain what is special about the failure implicated in doing the unthinkable, especially in comparison to the failure implicated in akratic action. The second is ‘Incapacity’: a theory of unthinkability must distinguish between not doing something because it is unthinkable and not doing something because of an extremely negative psychological reaction.

More and more, we face AI-based products and services. Using these services often requires our explicit consent, e.g., by agreeing to the services’ Terms and Conditions clause. Current advances introduce the ability of AI to evolve and change its own modus operandi over time in such a way that we cannot know, at the moment of consent, what it is in the future to which we are now agreeing. Therefore, informed consent is impossible regarding certain kinds of AI. Call this the problem of radical ignorance. Interestingly, radical ignorance exists in consent contexts other than AI, where it seems that individuals can provide informed consent. The article argues that radical ignorance can undermine informed consent in some contexts but not others because, under certain institutional, autonomy-protecting conditions, consent can be valid without being (perfectly) informed. By understanding these institutional conditions, we can formulate practical solutions to foster valid, albeit imperfectly informed consent across various decision contexts and within different institutions.

This paper develops Bernard Williams’s claim that moral incapacity – i.e., one’s inability to consider an action as one that could be performed intentionally – ‘is proof against reward’. It argues that we should re-construe the notion of moral incapacity in terms of self-identification with a project, commitment, value, etc. in a way that renders this project constitutive of one’s self-identity. This consists in one’s being insensitive to incentives to reconsider or get oneself to change one’s identification with this project. More precisely, self-identification with a project implies that no state-given reason can justify for oneself reconsidering, or getting oneself to revise, or abandon one’s identification with that project. This view ties together integrity and self-identification, and avoids problems common to competing views: it avoids regress problems faced by hierarchical theories of identification; it demonstrates that integrationist views of identification overlook the fact that a deep, well-integrated attitude may fail to be incentive-insensitive; and it helps explain what’s wrong with ‘perverse’ cases, where one values acting in a way that one does not all things consider value. It also improves on Williams’s own view, by construing moral incapacity not merely in terms of one’s incapacity to perform an action (that undermines one’s project and thus violates one’s integrity), but also in terms of one’s incapacity to reconsider one’s commitment (to said project).

In recent years, insurance companies have begun tracking their customers’ behaviors and price premiums accordingly. Based on the Market-Failures Approach as well as the Justice-Failures Approach, I provide an ethical analysis of the use of tracking technologies in the insurance industry. I focus on the use of telematics in car insurance and on the use of fitness tracking in life insurance. The use of tracking has some important benefits to policyholders and insurers alike: it reduces moral hazard and fraud, increases actuarial fairness, and incentivizes safe behavior. These benefits, however, are outweighed by significant moral objections. First, the use of tracking technologies significantly undermines the fairness of the interaction between insurance companies and policyholders. Specifically, the use of tracking eliminates information asymmetries, but in such a way that favors insurers exclusively. Furthermore, tracked behaviors and the ability to choose to behave safely are highly correlated with other variables such as income. Therefore, tracking-based insurance relies on and exacerbates existing inequalities and injustices, which undermines the benefits that tracking is supposed to introduce.

ABSTRACT:Technological advances introduce the possibility that, in the future, firms will be able to use big-data analysis to discover and offer consumers their individual reservation price. This can generate some interesting benefits, such as a better state of affairs in terms of equality of both welfare and resources, as well as increased social welfare. However, these benefits are countered by considerations of relational equality. This article takes up the market-failures approach as its basis to demonstrate what is wrong with using big data to personalize prices. The article offers an improvement to the market-failures approach and argues that what is wrong with using big data to personalize prices is that it unfairly undermines consumers’ ability to benefit from the market, which is the very point of having a market.

Joseph Heath (2014) argues that the contribution of competitive markets to Pareto-efficiency generates moral constraints that apply to business managers. Heath argues that ethical behavior on the part of management consists in avoiding profit-seeking strategies which, under conditions of perfect competition, would decrease Pareto-efficiency. I argue that because (1) such conditions do not obtain; and (2) the most efficient result – under imperfect conditions – is not achieved by satisfying the largest possible set of the remaining conditions; it is (3) impossible to draw any substantive ethical guidelines from Heath’s approach.