Ethics and AI | Part 2

Ethics for AI beginners. Business as usual or an inflexion point?

An incursion between various definitions of ethics does not help too much those who try to depict – in one single pronouncement – a clear landscape of those characteristics of AI which may make it prone to breaking ethical principles. Let us examine a sample of such definitions, picked randomly without pretence of exhaustiveness, and make some ad hoc conclusions on what seems to be relevant for normative purposes. 

Moral and morality are also institutional attributions, and so are ethics.

What is ethics for AI?

1. The study of the concepts involved in practical reasoning: good, right, duty, obligation, virtue, freedom, rationality, choice.¹

This is a classic, academic, all-encompassing definition, which appears to work for any kind of human undertaking, its basic being the Greek word ethos, which means, among others, character or characteristics. 

The term was originally used by Aristotle, as a person’s character or personality, in the broader context of rhetoric. In the famous triad, besides logos (logic, reason) and pathos (emotions), ethos would mean credibility and play the role of ensuring a balance between passion and caution.

In contemporary language, ethos has a different meaning and refers to the practices or values that distinguish one person, organization, or society from others, as manifested in their attitudes and values. Derived from that, ethics is: 

2. “A branch of philosophy dealing with what is morally right or wrong”.²

While the first definition seemed to indicate that ethics referred to an individual, with the second expansion of meaning we moved towards collective relevance (institutions, organizations, communities). Moral and morality are also institutional attributions, and so are ethics. 

This assumption is confirmed by the following definition, which extrapolate ethics to both individual and institutions.

3. An initial definition of ethics is the analysis, evaluation, and promotion of correct conduct and/or good character, according to the best available standards. […] Ethics asks what we should do in some circumstances, or what we should do as participants in some form of activity or profession. Ethics is not limited to the acts of a single person. Ethics is also interested in the correct practices of governments, corporations, professionals, and many other groups.³

This description also opens the way to the question of standards. Defining standards is not easy either, but it is done. Various alternatives prove the complexity of claiming the existence of a set of standards covering all requirements that would serve the concept of ethics.

4. An ethic is framework, or guiding principle, and it’s often moral. […]  A social ethic might include “treating people as you want to be treated.” Used in the plural, ethics refers to the moral rules that you live by.⁴ 

5. We can think of ethics as the principles that guide our behaviour toward making the best choices that contribute to the common good of all.⁵

6. Ethical theories study human moral behaviour and attempt to discover normative rules or maxims that describe what can be called “right action” and “wrong action.” Theories of ethics can be deontological systems, which are built around absolute moral rules that must be followed regardless of the outcome.⁶ 

7. Ethics is based on well-founded standards of right and wrong that prescribe what humans ought to do, usually in terms of rights, obligations, benefits to society, fairness, or specific virtues. 8. Ethics is the discipline concerned with what is morally good and bad, right and wrong.⁷

8. Ethics is the discipline concerned with what is morally good and bad, right and wrong.⁸

A simple enumeration of the notions that may accompany ethics – moral and immoral, right and wrong, good and bad, correct and incorrect – would be a clear indication that we are talking about human choices. Consequently, direct attribution of ethical features to any of the present AI systems is overstretched and misleading. Technology as such is supposed to be morally neutral.

Ethical concerns ultimately lie with humans, during all AI design, development, and deployment stages.

The simple notion of “AI ethics” is confusing. The imperative need in using AI systems and applications is human oversight. Ethics are the responsibility of humans wherever they are. This means that ethical concerns ultimately lie with humans, during all AI design, development and deployment stages. The right approach at this point should culminate with the ethical use of AI. 

There are, of course, the dangers and risks associated with autonomous decision-making by AI, where the human oversight will have the higher responsibility to ensure that ethical standards are upheld. History has shown us that humans will never be capable of mastering technology to ensure that it is used only for good and moral purposes. On the contrary, even at this stage an immeasurable amount of talent and creativity is spent in using AI for fake news, manipulation, propaganda, and plagiarism.

It is fair to note that international organisations, the United Nations in particular, contextualised the need for ethics whenever they dealt with international cooperation in the use of technologies. One such example is the World Summit on the Information Society and the Geneva Declaration of Principles.⁹ (see text box below).

In the sections to come, we shall try to identify what realistic measures international organizations could undertake to solve problems triggered not by AI as such, but by the humans who are using and abusing tools in their possession, including AI.

We do not have to rely on artificial intelligence to solve our old and new problems, we have to cultivate and strengthen human intelligence, including by using and improving technological tools. This leads us to a conclusion drawn by one of the most careful philosophers of ethics, Spinoza.

Without intelligence there is not rational life: and things are only good, in so far as they aid man in his enjoyment of the intellectual life, which is defined by intelligence. Contrariwise, whatsoever things hinder man’s perfecting of his reason, and capability to enjoy the rational life, are alone called evil.¹⁰

Now that we know what is old in ethics, let us see if and what the AI advent brings us in global governance.

[Part 2 of 6-part series]

---

1. Source: www.oxfordreference.com ↩︎
2. Source: www.merriam-webster.com/dictionary/ ↩︎
3. Ethics in a nutshell, University of Wisconsin-Madison, Centre for Journalism Ethics. ↩︎
4. Source: www.vocabulary.com ↩︎
5. Source: Santa Clara Institute, Ethics in Life and Business, www.scu.edu ↩︎
6. Source: Ethical Theories. Definitions & Examples, Study.com) ↩︎
7. Source: Markkula Centre for applied ethics), What is ethics, Manuel Velazquez, Clair Andre, Thomas Shank, Michael J. Meyer ↩︎
8. Britannica, Fifteenth edition, Micropaedia, 4:578:3a ↩︎
9. World Summit on the Information Society, Declaration of Principles, Building the Information Society: a global challenge in the new Millennium, Document WSIS-03/Geneva/Doc./4-E, December 2003. ↩︎
10. Spinoza, Ethics, Part IV. ↩︎

---

Dr Petru Dumitriu was a member of the Joint Inspection Unit (JIU) of the UN system and former ambassador of the Council of Europe to the United Nations Office at Geneva. He is the author of the JIU reports on ‘Knowledge Management in the United Nations System’, ‘The United Nations – Private Sector Partnership Arrangements in the Context of the 2030 Agenda’, ‘Strengthening Policy Research Uptake’, “Cloud Computing in the United Nations System”, and “Policies and Platforms in Support of Learning”. He received the Knowledge Management Award in 2017 and the Sustainable Development Award in 2019 for his reports. He is also the author of the Multilateral Diplomacy online course at DiploFoundation.