Details about the accuracy of the algorithm

NotAnonCo has developed a machine learning application to determine the identity of users who post anonymous comments (e.g., using pseudonyms or fake names) in public social media websites (e.g., YouTube, blog comments, etc). The algorithm compares the content of posts made with the same pseudonyms, across multiple social media websites, and compares the post content against public information (e.g., available from Facebook).

NotAnonCo promotes their service to various types of organisations for a fee. For example, some clients have used the service to check the suitability of new employees. Other clients use the service to identify customers who have posted negative reviews.

NotAnonCo does not provide any details about the accuracy of the algorithm. Some client organisations have received complaints after potential employees or customers have suffered consequences (such as being refused credit for home loans, not being considered for lucrative job offers, etc) after their identity was incorrectly linked to anonymous posts made by other people.

Who is morally responsible for the negative consequences suffered by people whose details were incorrectly linked to anonymous posts? Analyse each stakeholder’s responsibility and provide reasons to support your arguments.