The ethics of using AI addresses questions that legal regulations do not cover, but which determine trust within the company and among customers. This includes whether and when the use of a tool in producing an output is disclosed, how data on third parties that ends up in a prompt is handled, whether it is acceptable to create content imitating a specific person, how recommendations that could disadvantage someone are dealt with, and who bears responsibility when an output causes harm. For a company, it is more practical to have a few clear rules than an extensive code – for example, that the person who sends an output is responsible for it, and that sensitive data is not entered into tools outside an approved list. These rules belong in training, not only in a policy document.
See also: Internal AI usage policy, Verifying AI outputs, AI-generated Content.
