Algorithmic transparency is the principle that the factors that influence the decisions made by algorithms should be visible, or transparent, to the people who use, regulate, and are affected by systems that employ those algorithms [1].

Others [2], argue that it can serve multiple purposes:

1. Diakopoulos, N., & Koliska, M. (2017). Algorithmic transparency in the news media. Digital journalism, 5(7), 809-828.
2. Tal, A. S., Batsuren, K., Bogina, V., Giunchiglia, F., Hartman, A., Loizou, S. K., … & Otterbacher, J. (2019, June). “End to End” Towards a Framework for Reducing Biases and Promoting Transparency of Algorithmic Systems. In 2019 14th International Workshop on Semantic and Social Media Adaptation and Personalization (SMAP) (pp. 1-6). IEEE.

Based on the Ethics Guidelines of Trustworthy AI, the EU refers to Human agency highlighting the protection of individual [users’] autonomy which must be central to the system’s functionality. As they state, the key to this is the right not to be subject to a decision based solely on automated processing when this produces legal effects on users or similarly significantly affects them.

On the same line, human oversight helps ensure that an AI system does not undermine human autonomy or causes other adverse effects and it can be achieved using various governance mechanisms that involve human(s)-in-the-loop.

European Union (8 April 2019), Ethics Guidelines for Trustworthy AI. Retrieved 8 May 2021.

An algorithmic system can be considered as “fair” if similar people are being treated equally in the classification while still allowing a preferential treatment of individuals in the group. This approach can be used for certify fairness or for detecting unfairness in the system [1]. Fairness can be classified into two classes such as Individual Fairness and Group Fairness [2].

1. Dwork, C., Hardt, M., Pitassi, T., Reingold, O., & Zemel, R. (2012, January). Fairness through awareness. In Proceedings of the 3rd innovations in theoretical computer science conference (pp. 214-226).
2. Zemel, R., Wu, Y., Swersky, K., Pitassi, T., & Dwork, C. (2013, May). Learning fair representations. In International conference on machine learning (pp. 325-333). PMLR.

Algorithmic accountability implies that the organizations that use algorithms must be accountable for the decisions made by those algorithms, even though the decisions are being made by a machine, and not by a human being.

Dickey, Megan Rose (30 April 2017). Algorithmic Accountability. TechCrunch. Retrieved 8 May 2021.

Explainability is a multifaceted topic [1]. Explainable AI (XAI), for example, is artificial intelligence (AI) in which the results of the solution can be understood by humans. It contrasts with the concept of the “black box” in machine learning where even its designers cannot explain why an AI arrived at a specific decision [2].

1. Cliff Kuang (21 November 2017). Can A.I. Be Taught to Explain Itself?, The New York Times Magazine. Retrieved 8 May 2021.
2. Ian (5 November 2017). “Computer says no: why making AIs fair, accountable and transparent is crucial”. The Guardian. Retrieved 30 January 2018.

Scientists report six ethical concerns. Three of the ethical concerns refer to epistemic factors, specifically: inconclusive, inscrutable, and misguided evidence. Two are explicitly normative: unfair outcomes and transformative efects; while one—traceability—is relevant both for epistemic and normative purposes.

Tsamados, A., Aggarwal, N., Cowls, J., Morley, J., Roberts, H., Taddeo, M., & Floridi, L. (2021). The ethics of algorithms: key problems and solutions. AI & SOCIETY, 1-16.