AI Virtues As Missing Bedrock Ingredient For Responsible AI Says AI Ethics And AI Law

Are you virtuous?

Before you answer that question, let’s unpack the meaning of virtue and you can then take a concerted stab at elucidating your virtuousness.

Also, you’ll be perhaps surprised to know that virtues are a rising topic in the field of Artificial Intelligence (AI), especially in the realm of AI Ethics and AI Law. I’ll first cover some fundamentals about virtues and then jump into the latest thinking on AI Virtues. Yes, in short, AI Virtues are being bandied around as a precursor of sorts to Ethical AI and ultimately AI Law as well. For my ongoing and extensive overall coverage of AI Ethics and AI Law, see the link here and the link here, just to name a few.

First, ponder what is meant by referring to virtues and virtuousness.

Some would say that virtue is a quality or characteristic of showcasing outstanding morals. A virtuous person abides by the highest ethical or moral standards. The Latin word virtus was purportedly used by the Romans to emphasize moral rectitude and particularly valorous or heroic behavior. These are all rather high-minded concepts and crucial connotations associated with being virtuous.

There has been quite a lot of handwringing throughout history about what constitutes the core or keystones of virtue. Are there three keystones, five keystones, ten keystones, or how many exactly might there be?

For example, the so-called cardinal virtues are said to be these four precepts:

1) Prudence

2) Fortitude

3) Temperance

4) Justice (fairness)

Perhaps you should be mulling over your virtuousness by examining yourself in a serious sense of self-reflection on the basis of those four keystones. Do you showcase top-notch prudence? Do you showcase top-notch fortitude? Do you showcase top-notch temperance? And do you showcase top-notch fairness (justice) in your endeavors?

Notice that I keep saying that you have to “showcase” these foundational measures. It is one thing to perhaps harbor those in your mind, and it can be an entirely different matter to invoke those into real-world actions and deeds. The possibility of being virtuous solely in your mind is not what we are aiming to consider here. You have to take what is in your mind and turn it into reality.

Talk the talk, plus you need to genuinely walk the talk.

Another aspect is that you presumably need to abide by all of a considered set of keystones in order to truly be virtuous. If you were to be topnotch in prudence but failing or even slovenly in the other three of fortitude, temperance, and justice, you cannot seem to exhort that you are in fact virtuous. You are only partially so. We will insist that only if you abide by all such precepts can you carry proudly and loudly the flag of virtuousness. Thus, just meeting one, two, or three of these four precepts is insufficient. You might attain all four.

I am going to up the ante, so you’d better sit down for this next twist.

Researchers that have examined the list of virtues that seemed somewhat prominent during the Renaissance and other historical periods are apt to claim that there were seven virtues at that time:

1) Humility

2) Kindness

3) Temperance

4) Chastity

5) Patience

6) Charity

7) Diligence

If that lucky seven was indeed the case, and if you are going to use those as the stated criteria for virtuousness, I regret to inform you that the prior list of four required virtues is expanded into seven. This means that whereas you only had four to abide by earlier, now you’ve got a whopping seven to contend with.

Once again, start to do a self-reflection across those seven keystones.

I dare say the bar seems to be rising and rising. Maybe you at first assumed that of course you are virtuous, but now as the hurdles keep coming your way of these now seven keystones, it might be a lot harder to make a bold proclamation of your virtue-boasting prowess.

The number of keystones can get quite high.

Benjamin Franklin famously noted that he believed there were thirteen keystones for being virtuous, consisting of (as laid out in his autobiography):

1) Temperance

2) Silence

3) Order

4) Resolution

5) Frugality

6) Industry

7) Sincerity

8) Justice

9) Moderation

10) Cleanliness

11) Tranquility

12) Chastity

13) Humility

Yikes, that’s a daunting list.

Keep in mind too that we are laying down a gauntlet that asserts you must be all of those keystones and cannot be less than ideal on any of them. It is an all-or-nothing proposition. We might concede that you could be partially virtuous by abiding by some of the thirteen rather than all of them. We might also be willing to concede that you are partially virtuous if you at times are fully virtuous on all of them but then at other moments in time you are not attaining such completeness.

The gold star though only goes to those that attain all of the keystones at all times.

How many of us can meet that rigorous definition?

I guess that if you are raising your hand to have achieved this level of realization, we must do a cordial and congratulatory tip of the hat to you. Just to let you know, skeptics and cynics are bound to question the sincerity of your claims. Be so advised.

Now that we’ve got virtues onto the table, as it were, you might be wondering how virtues relate to ethics.

I’m glad that you asked.

One ardent belief is that virtues are what stoke ethics.

In this considered framework, your virtues will give rise to ethical efforts. Ergo, when someone rattles off a list of their ethical principles or rules, those are only rated as earnestly bona fide if also rooted in virtues. Your virtues drive toward your semblance of ethics.

I suppose a handy analogy might help.

We plant some seeds to grow flowers. The soil is essential to how those flowers are going to grow. We might be willing to say that the soil is the set of virtues, while the seeds and the flowers are the ethics that flow from that bedrock. Without suitable bedding, nothing else is likely to gain traction. You can chatter all day long about growing those flowers, but if the soil is bad or not conducive to matters at hand, you’ll need to anticipate faulty flowering or other related maladies will ensue.

You are welcome to make an objection to the crude analogy. Just wanted to get across an overarching sentiment on this virtue versus ethics conundrum. Be aware that some would say that virtues and ethics are one and the same. Others would say they are different. Of those that say they are different, there is the camp that indicates virtues are the bedrock and ethics are the outcropping (meanwhile, other camps with disparate views exist too).

To make it demonstrably apparent, not everyone agrees with the bedrock and outcropping way of looking at these matters. Nonetheless, we will proceed on that basis in this particular discussion. You can certainly argue at length at such a presumptive presumption, but at least be cognizant that this is an assumption herein and undertaken for the simple sake of fruitful discourse.

As a shorthand, let’s coin this as the virtue ethics conception.

Shifting gears a tad, mull over the trend today of flavorful ethics articulation and pontification.

A lot of debate takes place these days about ethics and asks openly what ethical rules or principles we are to observe. This is the case for how people behave. Furthermore, as you’ll see in a moment, there are proposed ethical rules or principles that AI ought to also observe.

The catch to the ruminations about ethical rules and ethical principles is that maybe we are concentrating on the wrong thing. Those in the virtue ethics camp would argue that rather than getting mired in the endless list of ethical rules and such, we would be wiser to focus on virtues. Get the virtues straightened out first, and from which the ethical rules will naturally flow.

Back to my analogy, we can get into heated discussions all day about the flowers we want to grow, but if we don’t first examine and make sure that we have the appropriate soil, none of the flower-growing ambitions are going to be of much practical use. Imagine that you’ve decided upon a bunch of flowers that seem quite compelling, yet it turns out that upon planting those seeds they never reach fruition because the soil wasn’t righted first.

You might even suggest that the constant din of what ethical rules to follow has become an undue obsession. It keeps getting hotter and more embroiled. In turn, we stray further and further away from the inner truth of having to put our eyes on virtues first.

In an insightful paper entitled Ethics and Virtue by researchers Manuel Velasquez, Claire Andre, Thomas Shanks, S.J., and Michael J. Meyer, they provide this noteworthy remark that we might be able to turn the corner on this obstinance: “Fortunately, this obsession with principles and rules has been recently challenged by several ethicists who argue that the emphasis on principles ignores a fundamental component of ethics — virtue. These ethicists point out that by focusing on what people should do or how people should act, the ‘moral principles approach’ neglects the more important issue—what people should be. In other words, the fundamental question of ethics is not ‘What should I do?’ but ‘What kind of person should I be?’ According to ‘virtue ethics’, there are certain ideals, such as excellence or dedication to the common good, toward which we should strive and which allow the full development of our humanity” (posted at the Markkula Center for Applied Ethics, Santa Clara University).

You might be vaguely aware that the AI realm has had its abundant share of proposals about AI Ethics principles and rules. You can readily find all manner and a number of pronounced Ethical AI precepts. They are a veritable dime a dozen, one might say.

The shocker, if you will, consists of the recent bubbling contestation that maybe, just maybe, we should be giving due attention to AI Virtues. Cease the crazy and furious finger-pointing about AI Ethics rules and principles, and instead cast your vision toward AI Virtues. If we can get the AI Virtues figured out, the rest will be easy-peasy (well, kind of, or at least more sensibly undertaken).

Before leaping into the AI Virtues topic, I’d like to first lay some essential foundation about AI and particularly AI Ethics and AI Law, doing so to make sure that the discussion will be contextually sensible.

The Rising Awareness Of Ethical AI And Also AI Law

The recent era of AI was initially viewed as being AI For Good, meaning that we could use AI for the betterment of humanity. On the heels of AI For Good came the realization that we are also immersed in AI For Bad. This includes AI that is devised or self-altered into being discriminatory and makes computational choices imbuing undue biases. Sometimes the AI is built that way, while in other instances it veers into that untoward territory.

I want to make abundantly sure that we are on the same page about the nature of today’s AI.

There isn’t any AI today that is sentient. We don’t have this. We don’t know if sentient AI will be possible. Nobody can aptly predict whether we will attain sentient AI, nor whether sentient AI will somehow miraculously spontaneously arise in a form of computational cognitive supernova (usually referred to as the singularity, see my coverage at the link here).

The type of AI that I am focusing on consists of the non-sentient AI that we have today. If we wanted to wildly speculate about sentient AI, this discussion could go in a radically different direction. A sentient AI would supposedly be of human quality. You would need to consider that the sentient AI is the cognitive equivalent of a human. More so, since some speculate we might have super-intelligent AI, it is conceivable that such AI could end up being smarter than humans (for my exploration of super-intelligent AI as a possibility, see the coverage here).

I’d strongly suggest that we keep things down to earth and consider today’s computational non-sentient AI.

Realize that today’s AI is not able to “think” in any fashion on par with human thinking. When you interact with Alexa or Siri, the conversational capacities might seem akin to human capacities, but the reality is that it is computational and lacks human cognition. The latest era of AI has made extensive use of Machine Learning (ML) and Deep Learning (DL), which leverage computational pattern matching. This has led to AI systems that have the appearance of human-like proclivities. Meanwhile, there isn’t any AI today that has a semblance of common sense and nor has any of the cognitive wonderment of robust human thinking.

Be very careful of anthropomorphizing today’s AI.

ML/DL is a form of computational pattern matching. The usual approach is that you assemble data about a decision-making task. You feed the data into the ML/DL computer models. Those models seek to find mathematical patterns. After finding such patterns, if so found, the AI system then will use those patterns when encountering new data. Upon the presentation of new data, the patterns based on the “old” or historical data are applied to render a current decision.

I think you can guess where this is heading. If humans that have been making the patterned upon decisions have been incorporating untoward biases, the odds are that the data reflects this in subtle but significant ways. Machine Learning or Deep Learning computational pattern matching will simply try to mathematically mimic the data accordingly. There is no semblance of common sense or other sentient aspects of AI-crafted modeling per se.

Furthermore, the AI developers might not realize what is going on either. The arcane mathematics in the ML/DL might make it difficult to ferret out the now-hidden biases. You would rightfully hope and expect that the AI developers would test for the potentially buried biases, though this is trickier than it might seem. A solid chance exists that even with relatively extensive testing that there will be biases still embedded within the pattern-matching models of the ML/DL.

You could somewhat use the famous or infamous adage of garbage-in garbage-out. The thing is, this is more akin to biases-in that insidiously get infused as biases submerged within the AI. The algorithm decision-making (ADM) of AI axiomatically becomes laden with inequities.

Not good.

All of this has notably significant AI Ethics implications and offers a handy window into lessons learned (even before all the lessons happen) when it comes to trying to legislate AI.

Besides employing AI Ethics precepts in general, there is a corresponding question of whether we should have laws to govern various uses of AI. New laws are being bandied around at the federal, state, and local levels that concern the range and nature of how AI should be devised. The effort to draft and enact such laws is a gradual one. AI Ethics serves as a considered stopgap, at the very least, and will almost certainly to some degree be directly incorporated into those new laws.

Be aware that some adamantly argue that we do not need new laws that cover AI and that our existing laws are sufficient. They forewarn that if we do enact some of these AI laws, we will be killing the golden goose by clamping down on advances in AI that proffer immense societal advantages.

In prior columns, I’ve covered the various national and international efforts to craft and enact laws regulating AI, see the link here, for example. I have also covered the various AI Ethics principles and guidelines that various nations have identified and adopted, including for example the United Nations effort such as the UNESCO set of AI Ethics that nearly 200 countries adopted, see the link here.

Here’s a helpful keystone list of Ethical AI criteria or characteristics regarding AI systems that I’ve previously closely explored:

• Transparency
• Justice & Fairness
• Non-Maleficence
• Responsibility
• Privacy
• Beneficence
• Freedom & Autonomy
• Trust
• Sustainability
• Dignity
• Solidarity

Those AI Ethics principles are earnestly supposed to be utilized by AI developers, along with those that manage AI development efforts, and even those that ultimately field and perform upkeep on AI systems.

All stakeholders throughout the entire AI life cycle of development and usage are considered within the scope of abiding by the being-established norms of Ethical AI. This is an important highlight since the usual assumption is that “only coders” or those that program the AI is subject to adhering to the AI Ethics notions. As prior emphasized herein, it takes a village to devise and field AI, and for which the entire village has to be versed in and abide by AI Ethics precepts.

I also recently examined the AI Bill of Rights which is the official title of the U.S. government official document entitled “Blueprint for an AI Bill of Rights: Making Automated Systems Work for the American People” that was the result of a year-long effort by the Office of Science and Technology Policy (OSTP). The OSTP is a federal entity that serves to advise the American President and the US Executive Office on various technological, scientific, and engineering aspects of national importance. In that sense, you can say that this AI Bill of Rights is a document approved by and endorsed by the existing U.S. White House.

In the AI Bill of Rights, there are five keystone categories:

• Safe and effective systems
• Algorithmic discrimination protections
• Data privacy
• Notice and explanation
• Human alternatives, consideration, and fallback

I’ve carefully reviewed those precepts, see the link here.

Now that I’ve laid a helpful foundation on these related AI Ethics and AI Law topics, we are ready to jump into the heady topic of AI Virtues.

AI Virtues Getting Tuned Up To Aid Ethical AI

When I refer to AI Virtues, please realize that I am not talking about sentient AI.

If (or some would contend when) we reach sentient AI, the sentient AI might or might not have a set of virtues. We can argue until the cows come home as to whether or not sentient AI will include its own semblance of virtues. Some pundits would insist that virtues are a humankind-only element that will be entirely and unquestionably absent from sentient AI.

Other pundits proclaim the exact opposite, namely that sentient AI will of course have virtues. In that latter case, we ought to be worrying about how to ensure that sentient AI has the right kind of virtues. Perhaps we can feed our virtues into the AI that becomes sentient. If that won’t work, we are going to hope that sentient AI is clever enough to realize the importance of virtues and divine them of its own volition.

Round and round that merry-go-round goes.

For today’s viewpoint, I’d like to keep our eyes on contemporary non-sentient AI.

In a provocative research paper entitled “A Virtue‑Based Framework To Support Putting AI Ethics Into Practice,” author Thilo Hagendorff proposes that we utilize virtue ethics and as a result might arrive at the belief that there are four foundational AI Virtues:

1) Justice

2) Honesty

3) Responsibility

4) Care

The paper posits that “Many ethics initiatives have stipulated sets of principles and standards for good technology development in the AI sector. However, several AI ethics researchers have pointed out a lack of practical realization of these principles. Following that, AI ethics underwent a practical turn, but without deviating from the principled approach. This paper proposes a complementary to the principled approach that is based on virtue ethics. It defines four “basic AI virtues”, namely justice, honesty, responsibility and care, all of which represent specific motivational settings that constitute the very precondition for ethical decision making in the AI field” (in Philosophy & Technology, June 2022).

How did the four AI Virtues get derived?

According to the researcher, it is feasible to examine the multitude of AI Ethics precepts and pretty much back your way into what the underlying keystone AI Virtues would necessarily be. All you need to do is a bit of mindful analysis and you can turn the morass into something nice and neat. As stated: “When sifting through all these principles, one can, by using a reductionist approach and clustering them into groups, distill four basic virtues that cover all of them” (ibid).

Various charts and figures are included, which proffer that we might construe the four AI Virtues as consisting of these embodiments when it comes to AI Ethics principles or rules:

• AI Virtue of Justice: AI Ethics precepts include algorithmic fairness, non-discrimination, bias mitigation, inclusion, equality, diversity, etc.
• AI Virtue of Honesty: AI Ethics precepts include transparency, openness, explainability, interpretability, technological disclosure, open source, acknowledge errors and mistakes, etc.
• AI Virtue of Responsibility: AI Ethics precepts include responsibility, liability, accountability, replicability, legality, accuracy, considering long-term technological consequences, etc.
• AI Virtue of Care: AI Ethics precepts include non-maleficence, harm, safety, privacy, protection, precaution, hidden costs, beneficence, well-being, sustainability, peace, common good, solidarity, social cohesion, freedom, autonomy, liberty, consent, etc.

Putting yourself into the shoes of having to do this kind of reverse engineering, this is what the research paper suggested ought to be mentally performed: “Does virtue A describe character dispositions that, when internalized by AI practitioners, will intrinsically motivate them to act in a way that ‘automatically’ ensures or makes it more likely that the outcomes of their actions, among others, result in technological artefacts that meet the requirements that principle X specifies? Or, in short, does virtue A translate into behavior that is likely to result in an outcome that corresponds to the requirements of principle X?” (ibid).

It is probably safe to say that we might not all reach the same conclusions.

There would seem to be plenty of room to argue that a particular AI Ethics precept belongs in some particular AI Virtue or reasonably placed into some other one, or maybe belongs in more than one, etc. This can readily go back and forth, doing so in a civilized and well-mannered tone (no need to switch into angry polarizing angsts).

You can go even deeper into this slicing-and-dicing by coming up with additional AI Virtues beyond the stated four and making a reasoned claim that there are more AI Virtues to be had. One supposes you could also try to reduce the count to say only three or two AI Virtues, though that would likely put you on a rather wobbly philosophical and semi-impractical grounding.

Before you start over-analyzing the four AI Virtues, you ought to know that the research paper indicates that there are two additional second-order AI Virtues that come into the mix. The two added second-order AI Virtues are:

Those seem to consist of:

• AI Virtue of Prudence: AI Ethics precepts entailing System 1 thinking, implicit biases, in-group favoritism, self-serving biases, value-action gaps, moral disengagement, etc.
• AI Virtue of Fortitude: AI Ethics precepts entailing situational forces, peer influences, authorities, etc.

The rationale or basis for these two second-order AI Virtues is partially based on this notion: “While both virtues may help to overcome bounded ethicality, they are at the same time enablers for living up to the basic virtues. Individual psychological biases as well as situational forces can get in the way of acting justly, honestly, responsibly or caringly. Prudence and fortitude are the answers to the many forces that may restrict basic AI virtues, where prudence is aiming primarily at individual factors, while fortitude addresses supra-individual issues that can impair ethical decision making in AI research and development” (ibid).

All told, if I can try to recap the proposed set of AI Virtues, they are this:

• Justice
• Honesty
• Responsibility
• Care
• Prudence (second-order)
• Fortitude (second-order)

Which generally tend to consist of this coupling with AI Ethics rules or principles:

• AI Virtue of Justice: AI Ethics precepts include algorithmic fairness, non-discrimination, bias mitigation, inclusion, equality, diversity, etc.
• AI Virtue of Honesty: AI Ethics precepts include transparency, openness, explainability, interpretability, technological disclosure, open source, acknowledge errors and mistakes, etc.
• AI Virtue of Responsibility: AI Ethics precepts include responsibility, liability, accountability, replicability, legality, accuracy, considering long-term technological consequences, etc.
• AI Virtue of Care: AI Ethics precepts include non-maleficence, harm, safety, privacy, protection, precaution, hidden costs, beneficence, well-being, sustainability, peace, common good, solidarity, social cohesion, freedom, autonomy, liberty, consent, etc.
• AI Virtue of Prudence (second-order): AI Ethics precepts entailing System 1 thinking, implicit biases, in-group favoritism, self-serving biases, value-action gaps, moral disengagement, etc.
• AI Virtue of Fortitude (second-order): AI Ethics precepts entailing situational forces, peer influences, authorities, etc.

A spirited debate over all of this is easy to be prodded into action.

I’m sure that some of you right now are already going berserk over one or another of the proffered AI Virtues. It’s not a virtue, some might be yelling. It is not adequately worded some of you might be crying out. Quite a shouting match could arise.

Amidst the various anticipated criticism and qualms that the research explores, one that seems to me especially notable has to do with the classic agent-centered versus act-centered issues involved. The agent-centric viewpoint is basically that we want an agent or actor to be of a certain mind, while the act-centric viewpoint tends to focus on the actions that are undertaken.

Perhaps, some would say, AI Virtues are more so about the agency or agent-centric aspects, while AI Ethics is more so about the act-centric constituents. We want AI developers and AI systems to be rooted in AI Virtues as a type of “mindset” (programming in the case of the AI), and the acts of the AI developers and the acts of the AI are represented via the Ethical AI precepts.

There is also the danger that some will misrepresent this as though the use of AI Virtues implies that AI will never do any wrong. Or perhaps we will get bogged down in devising AI Virtues and meanwhile lose sight of AI Ethics principles or rules. Plenty of motivations can be conjured to undercut the AI Virtues as a framework that either doesn’t do much good or worse so distracts and confounds the work that a displaced commentator might holler really needs to be done instead.

Conclusion

Do we need AI Virtues?

And, if so, will they be of use and warmly embraced by those already steeped in AI Ethics, or might AI Virtues be seen as duplicative, a false decoy, an annoying distraction, or otherwise portrayed as something alluring but an unsuitable splintering of already overwhelmed efforts to get Ethical AI into the minds and hands of companies devising and using AI.

Many AI Ethics advocates already are overtaxed as to getting business leaders to listen and getting AI adopters to seriously consider Ethical AI precepts (for my coverage of AI Ethics Boards, see the link here, and for analysis of worker burnout for those carrying the banner of Ethical AI, see the link here). You can envision that a common reaction to AI Virtues would be that the plate is already full of AI Ethics rules, thus, let’s get those fully consumed before we venture into the stratosphere of AI Virtues.

A counterargument is that we inadvertently skipped over or neglected to go to basics. AI Virtues should have long ago been laid out. Though we cannot turn back the clock, we can try to make up for the lost time. The same mighty mechanizations over AI Ethics can certainly absorb the belated inclusion of AI Virtues.

Get to it, stop complaining.

Friedrich Nietzsche said: “We do not place special value on the possession of a virtue until we notice its total absence in our opponent.” You might then agree that the advent of AI For Bad has sparked not only our realization of the need for AI Ethics but has also likewise (or ought to have) stoked our appetite for and possible acceptance of AI Virtues too.

The last word on this goes to Marcus Tullius Cicero, Roman statesman, for which he reportedly exclaimed: “The enemy is within the gates; it is with our own luxury, our own folly, our own criminality that we have to contend.” You see, the vices are already within the gates of AI, and we might need to up our attention to virtues to counteract the rising tide of evils.

AI Virtues are waiting patiently but persistently at the gates.