ABSTRACT
This chapter discusses the permissible and impermissible use of software agents to influence internet user behaviour through targeted content such as app notifications, product recommendations, and prices for goods and services. It is argued that impermissible behavioural influence typically involves manipulation, coercion, or deception. However, existing analyses of what explains the wrongness of manipulation, coercion, and deception make ineliminable reference to mental state terms such as intention, which software agents cannot obviously satisfy. A novel account is developed of what explains the wrongness of manipulation, coercion, and deception, on the part of software agents. On this view, morally impermissible instances of behavioural influence by software agents are wrong in virtue of undermining the mental integrity of internet users, which is to say, roughly, that these influence strategies target certain preconditions for authentic decision-making on the part of users.
