I was once asked whether it is possible to embed character into synthetic humans. For those unfamiliar with “synths,” as they are often referred to, Cathy Hackl, a contributor to Forbes, provides an excellent overview of these human-like robots designed to be indistinguishable from real humans. Synths are engineered to go beyond generative artificial intelligence (AI) to artificial general intelligence, with the added benefit of their human-like design, which seeks to replicate how humans process information, learn, and interact. For example, as Jun Wu, another Forbes contributor, explains, empathy—an attribute often missing in AI—is crucial for artificial general intelligence and an essential characteristic of synths. Imagining AI-powered synths is the ultimate test for character since many critiques of AI zero in on its lack of human qualities.

Contemplating the possibility of embedding character in synths leads us to Shannon Vallor’s argument in her book, the AI Mirror. Vallor warns that AI, like a powerful but flawed mirror, reflects the biases and failures of the wisdom humans strive to escape. To truly comprehend AI’s flawed mirror, we must first understand our own character. By examining how character is revealed in humans and AI, we can strengthen the character-based judgment underpinning both.

A Primer On Character

The study of character, a discipline with a rich history dating back to influential figures like Confucius, Aristotle, and Plato, has recently gained traction in management. This renewed focus on character underscores its potential to enhance well-being and decision-making alongside competence for sustained excellence.

We developed the leader character framework that connects the science and practice of character, revealing 11 interconnected dimensions, with judgment—Aristotle’s “practical wisdom”—at the center (see Figure 1). Each dimension includes observable, assessable, and developable virtuous behaviors, which can exist in what are viewed as deficient or excess vice states. Excess occurs when a strength is unsupported by other necessary behaviors. For instance, courage can lead to recklessness without temperance or result in cowardice when deficient.

Character is the result of habits formed through a person’s life experiences. For example, every action we take can either strengthen or weaken our courage. When assessing someone’s character, it’s essential to consider how likely they are to demonstrate certain behaviors consistently, avoiding extremes that could indicate vice. Character development relies on four interconnected systems: physiology, affect, behavior, and cognition. For example, a person might intend to be courageous (cognition), but fear (affect) and a fight-or-flight response (physiology) can hinder action (behavior). While a person’s character can be influenced by their environment, strong character is less dependent on external factors, such as social norms and incentives, that might otherwise shape their behavior.

AI and Character

As Forbes contributor Craig Smith describes, artificial general intelligence can now be viewed on a spectrum with artificial superintelligence in reach. Although I refer to AI more broadly, we need to acknowledge AI exists along a spectrum, and the assessment of how character can or cannot be embedded in AI depends on the spectrum, with the added challenge that the spectrum is constantly shifting.

There are two critical points to consider as I examine how AI aligns with the 11 character dimensions and their associated behaviors. The first point is that weak or unbalanced character can lead to poor judgment. This reflects Vallor’s warning that AI can act as a flawed mirror when it reflects back to us the weak or unbalanced character of humans. The second point is that character cannot replace competence, and vice versa. Competence, which focuses on knowledge, skills, and abilities, presents a different challenge for AI. Regardless of whether AI can adequately substitute for competence, it cannot replace character.

Transcendence

The dimension of transcendence with its behaviors of being appreciative, inspired, purposeful, optimistic, and creative, points to its importance in cultivating a sense of possibility with clarity about what is important. Generative AI is designed to create new content and augment human creativity, as described by Eapen and colleagues in their 2023 Harvard Business Review article. They argue that it augments human creativity by promoting divergent thinking, challenging expertise bias, assisting in idea evaluation, supporting idea refinement, and facilitating collaboration. The key point is that Generative AI augments and is not a substitute for human creativity, and the same idea applies to the other behaviors of transcendence. Developing transcendence cultivates a way of being that fosters an expansive perspective in humans. Deficiencies in transcendence compromise judgment because they limit the possibilities people imagine. However, when strong transcendence is not supported by other dimensions people can risk being seen as delusional or untethered.

Drive

Although AI can be programmed to sound passionate, the gap between sounding and being is most apparent when thinking about it for humans. The challenges anyone experiences about becoming passionate and sustaining it provide insights into what is missing when we only focus on sounding passionate. Although results-oriented could be nicely AI-aligned, consider that striving for excellence requires an imagination for what excellence is, which must also be connected to the other dimensions of character, including justice and humanity so that drive does not operate in its excess state.

Collaboration

While AI can certainly foster collaboration, as described by Forbes contributor Peter Pluim, it’s important not to overlook the human underpinnings of character in this context. Becoming someone who is cooperative, collegial, open-minded, flexible, and interconnected is what truly enables effective engagement with others and AI. This emphasis on character underscores the human element in collaboration, exposing not only what can set humans apart from AI, but exposing the challenges humans face in cultivating the behaviors associated with collaboration.

Humanity

Humanity is one of the dimensions on which critiques of AI are mounted. The basic fear is that algorithms and computations will obscure humanity. Vallor talks about the danger of treating AI as if it were human. The rise of people seeking artificial intimacy through AI, as described by Forbes contributor Virginie Berger, reveals that AI can often seem more compassionate than humans. Arguably, AI might do a better job of forgiving and appearing considerate, and it is conceivable to program some protocols that humans often struggle with. However, at its best, strength in humanity will be unparalleled by AI. Yet the caution is that strong humanity, not supported by the other character dimensions, will compromise well-being and judgment, as seen in cases of burnout, compassion fatigue, or the inability to hold people accountable because a person doesn’t want to hurt others.

Humility

Although people often think of humility in its excess vice state of being self-conscious or self-effacing, they miss that high levels of humility, when supported by the other dimensions of character, are a powerhouse for learning and development. In contrast, a basic level of humility is built into generative AI. We don’t think of AI as struggling with being modest, vulnerable, respectful, or self-aware, whereas these are behaviors that humans struggle to develop. That these behaviors are not something we consider a part of AI suggests we need to contemplate that the deficient states of humility such as being unreflective or disinterested could be blind spots for AI. When developed in humans, strength in humility provides a distinctiveness that sets humans apart.

Integrity

Whereas humans struggle with candor, transparency, and consistency, AI could be characterized as inherently candid, transparent, and consistent. However, because AI has weaknesses in other character dimensions, these characteristics effectively operate in the excess vice state. For example, when consistency is not tied to being socially responsible, it becomes rigid, and candor unsupported by empathy becomes belligerent. AI lacks the principles and authenticity that reside within the person and are cultivated over time through lived experience. For humans, developing integrity will be an essential cornerstone for navigating how we engage with AI, ensuring we don’t lose our authenticity and principles. How our context shapes us, particularly social media, was flagged as an issue long before AI.

Temperance

Temperance has consistently been one of the weakest dimensions of character for most people. Because temperance has such clear connections to physiology and affect, it operates as a first line of defense in daily interactions. We notice quickly when we become impatient, for example. Although we don’t use terms such as being calm or patient when thinking of AI, it is instructive to think of AI as having an abundance of calm, because then we begin to understand it in the excess vice state of being indifferent. Often our patience and calm is tested because we are not indifferent owing to strengths in other dimensions such as justice or humanity that register fear or concern within us.

Justice

Justice is the dimension that many critics of AI are concerned about, and it is also where many issues around morals and ethics emerge. For example, because generative AI relies on what might be viewed as the collective consciousness of the world played out in words, images, and algorithms, it will embed the world’s injustices within it. Because justice is also connected to all the other character dimensions, our sense of what is fair or equitable is shaped by our humanity, for example. A person who lacks compassion or empathy will also find they have blind spots associated with justice. It isn’t simply about justice, but how justice is influenced by all character dimensions.

Accountability

Becoming someone who takes ownership, accepts consequences, is conscientious, and responsible, is challenging. How these characteristics are embedded in AI is questionable, and the same concerns raised in the prior section on justice apply to accountability. On what basis is AI accountable, and to whom? Developing accountability means also being accountable for the systems we design and rely on, including AI.

Courage

Becoming brave, determined, tenacious, resilient, and confident is arduous for humans. As with all dimensions of character, but seen clearly with this dimension, our life experiences shape us, and with courage, often through crucible moments. Although it is tempting to consider that AI has unlimited determination and tenacity with the ultimate resilience, as it operates for as long as it is programmed, such a perspective misses the essence of how human courage shapes judgment and is subject to the same concerns I raised about temperance. Courage is not simply a matter of the dogged determination arising from an algorithm, for example, but rather the life experiences that cultivate courage – what matters, what is complex, and how we bounce forward. All of which influence judgment.

Judgment

While there are many behaviors associated with judgment that AI can augment, such as being analytical and cognitively complex, several behaviors provide trump cards for character. The first is the set of behaviors that shape the other behaviors, such as being situationally aware, intuitive, and insightful. Also, because judgment acts as an air traffic controller for all character dimensions, it plays the pivotal role of cultivating the practical wisdom that Aristotle noted as critically important. It takes character-based judgment to use AI effectively and to exercise the ultimate judgment associated with the quality and merits of what AI produces.

Overall, understanding character reveals that humans have a long way to go to strengthen character. While some aspects of AI are character-aligned, there are a host of character deficiencies. These character deficiencies go beyond the usual suspects of justice and humanity. This basic overview reveals that much of what makes us human lies in all dimensions of character, which are cultivated through our life experiences. Every effort should be made to instill character into AI applications. However, we also need more clarity around the limitations of AI when it comes to character. The bottom line is that AI is a powerful tool that needs to be harnessed with a stronger underpinning of character-based judgment. For that, we need to start with strengthening our own character.



Source link

author-sign