A Unified Framework of Five Principles for AI in Society

By Luciano Floridi, Josh Cowls. 2019

1,532 citations recorded by the source corpus.

Artificial Intelligence (AI) is already having a major impact on society. As a result, many organizations have launched a wide range of initiatives to establish ethical principles for the adoption of socially beneficial AI. Unfortunately, the sheer volume of proposed principles threatens to overwhelm and confuse. How might this problem of ‘principle proliferation’ be solved? In this paper, we report the results of a fine-grained analysis of several of the highest-profile sets of ethical principles for AI. We assess whether these principles converge upon a set of agreed-upon principles, or diverge, with significant disagreement over what constitutes ‘ethical AI.’ Our analysis finds a high degree of overlap among the sets of principles we analyze. We then identify an overarching framework consisting of five core principles for ethical AI. Four of them are core principles commonly used in bioethics: beneficence, non-maleficence, autonomy, and justice. On the basis of our comparative analysis, we argue that a new principle is needed in addition: explicability, understood as incorporating both the epistemological sense of intelligibility (as an answer to the question ‘how does it work?’) and in the ethical sense of accountability (as an answer to the question: ‘who is responsible for the way it works?’). In the ensuing discussion, we note the limitations and assess the implications of this ethical framework for future efforts to create laws, rules, technical standards, and best practices for ethical AI in a wide range of contexts.KeywordsAccountability; Autonomy; Artificial Intelligence; Beneficence; Ethics; Explicability; Fairness; Intelligibility; Justice; Non-maleficence.

Subfields: Safety Research

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Safety Research

Citations in the corpus

Ethical principles in machine learning and artificial intelligence: cases from the field and possible ways forward (2020, background): “According to Floridi and Cowls (2019) a prominent role is also played by the autonomy dimension; the possibility of refraining from ceding decision power to AI for overriding reasons (e.g., the gain of ef ﬁ cacy is not deemed ﬁ t to justify the loss of control over decision-making).”
Ethical and regulatory challenges of AI technologies in healthcare: A narrative review (2024, background): “It appears that the concept of patients seeking a doctor’s expert advice and placing themselves under their care is evolving into a model where patients actively participate in generating health knowledge and acquiring expertise to manage their illnesses [85].”
Trustworthy Artificial Intelligence: A Review (2022, background): “Over the past few years, research institutes, private organizations, and government agencies have proposed various guidelines and frameworks to make AI trustworthy [60, 68, 69, 154, 198, 224].”
A comprehensive AI policy education framework for university teaching and learning (2023, background): “Floridi (2021)’s framework for the ethical use of AI, which proposed the 5 core principles of “beneficence, non-maleficence, autonomy, justice and explicability”, is referred to by most national policies on AI as a foundation to further develop on.”
How to Design AI for Social Good: Seven Essential Factors (2020, background): “The seven factors analysed in the previous pages are summarised in Table 1, together with the corresponding best practices, and the five principle(s) of AI ethics identified in (Floridi and Cowls 2019) to which each factor is most closely identified.”
The ethics of algorithms: key problems and solutions (2020, background): “(Floridi 2019b, 187).”
Trustworthy AI: From Principles to Practices (2021, background): “A number of enterprises [57, 62, 136, 166, 254, 338], academia [122, 199, 218, 301, 322], public sectors, and organizations [9, 210, 334] have recently identified these factors and summarized them as principles of AI trustworthiness.”
The uselessness of AI ethics (2022, background): “Despite the mush-rooming of ethical statements, Floridi and Cowls [21] suggest many values recur frequently and can be condensed into five core principles: beneficence, non-maleficence, autonomy, justice, and explicability.”
Trustworthy artificial intelligence (2020, background): “Toward this end, “ [one] must be able to understand the good or harm [AI] is actually doing to society, and in which ways ” (Floridi and Cowls 2019) for it to be beneficent and non-maleficent.”
From What to How: An Initial Review of Publicly Available AI Ethics Tools, Methods and Research to Translate Principles into Practices (2019, methodology): “It is, therefore, a necessary building block in the creation of an environment that fosters ethical, responsible, and beneficial ML, especially as it also indicates the possibility of a time when the distractive risk of ethics shopping 13 (Floridi 2019c) will be lessened.”
Education for AI, not AI for Education: The Role of Education and Ethics in National AI Policy Strategies (2021, methodology): “AIED scholars who wish to better incorporate ethical thinking into their work, such as by reference to the ethical framework of Floridi et al. (2019) explored in this article, have many sources on which to draw.”
Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback (2023, background): “Some works have begun to assess whether LLMs can be used to facilitate agreement between different humans [Bakker et al., 2022] and to codify the broad-ranging principles under which deployment of AI systems for public good can be assessed [Floridi and Cowls, 2022, Sartori and Theodorou, 2022].”