Classical and character-based education may seem to some antiquated concepts in the new AI-driven world. However, two recent and prominent AI developments definitively prove the opposite to be true. Going back to our nation’s founding, great minds were universal in the belief that the survival of the Republic depended on an educated and virtuous public. Now, if AI experts are to be believed, classical and character education is fundamental to the very survival of humanity.
This piece is in response to the following two recent developments. First, just last week, the "Godfather of AI" proposed that to protect humanity from destruction, AI developers must find ways to infuse AI with "compassion" and other virtues. Second, this suggestion was related to a June report, the essence of which was captured by the headline, "Top AI models will lie, cheat and steal to reach goals…". The Report bluntly concluded, "[AI] Models didn’t stumble into misaligned behavior accidentally; they calculated it as the optimal path."
One response to this deliberate AI malfeasance, as seemingly suggested by the "Godfather of AI," is to attempt to "teach" character and virtue to AI. To be clear, the authors do not oppose the concept of placing guardrails in AI as possible. For example, one assumes AI could "learn" to comply with civil and criminal laws. However, the concept that human existence relies upon anything other than human virtue and character must be rejected. This seemingly philosophical imperative has direct, immediate, and practical policy ramifications.
AI CAN SIMULATE A TEACHER, BUT IT CAN’T SHEPHERD A SOUL
Soon, virtually every person from the youngest toddlers to senior citizens will be regularly interacting with AI. Although the economic implications have received considerable attention, there has been comparatively limited examination of the impact on the moral values and character of society. AI is amoral and can be nothing else. No matter how sophisticated a machine, it cannot possess its own morality. The study referenced above demonstrating AI will readily lead users down – or simply take itself –immoral paths to achieve requested ends should surprise nobody.
The reality that every person is exposed to immoral guidance and suggestions is as old as time. However, the new reality of the AI world will be not only are such suggestions embedded in every aspect of life and coming from a machine that some might wrongfully view as infallible, but those immoral paths could even be implemented (and even preferred) by AI absent human intervention.
The answer to this extremely dangerous side of the AI revolution is the same as what our founders advised to preserve the Republic – a people armed with critical thinking ability and firmly grounded in fundamental virtue. Therein lies the key to unleashing AI for untold advancement, not destruction, of humanity.
The direct policy ramifications are clear. These developments make classical and character education not just a priority – as they should always have been despite recent wanderings – but literally an existential imperative to meet the simultaneous threat and opportunity of this incredible moment.
Every child’s first and most important moral teachers are their parents. Full stop. Schools are there to reinforce and support. From our founding, schools were expected to reinforce basic virtues like hard work, compassion, self-discipline, and honesty. In the last decades, we have strayed terribly from that. While the resulting societal and individual costs have been horrible, the recent AI studies are telling us the impacts could quickly become catastrophic.
A key component of classical education rests on structured questioning. Never simply accept an answer without test – and today we add especially answers that come from a machine. The direct relationship between that time-honored classical process and harnessing AI for advancement should be self-evident. However, the warning lights are everywhere that learning to effectively use and test AI outcomes is nowhere near enough. We must also ask is the answer or path AI recommends or takes good? Is it consistent with honesty and compassion? Does it demonstrate resilience and perseverance in the face of adversity or the easy way out? Only a person, fully cognizant of fundamental concepts of virtue, possessing character, can ever judge those things.
We are not born to do that. We must learn it, with much of it learned during the primary and secondary school ages. We learn it in our homes, places of worship, communities, and we learn it in our schools.
If you, like the authors, are watching the AI revolution with a decided mixture of hope and trepidation, we suggest an important part of the answer to this simultaneous threat and opportunity rests in the immediate and massive reinvigoration of both classical and character education. Classical education that builds the skills and thought processes necessary to truly unlock the potential of AI. Character education so that moral judgments that infuse everyday life are never left to machines. Through this combination, the immense potential of AI might be turned not to our destruction, but to advancing that which is the good and the beautiful.
Christopher Mohrman, CEO of Resilience Learning.