Introduction:

With all the buzz around Artificial Intelligence these days, and the endless debates about AI cheating and what’s right and wrong, how do we forge a path forward?

What we know is that AI is advancing in leaps and bounds. When I was completing my associate's degree from 2014 to 2017, AI was a distant dream. A professor of mine said that in his day, he dreamt of flying cars but could never predict the cell phone. Likewise, when I was playing on my Commodore 64 at 10 years old, my wildest technology dreams only included computers that could play music. I never dreamt of AI.

Now AI is everywhere, and every major corporation and government has their hand in the pot. AI is getting smarter at an unprecedented rate. So how do we ensure that as AI systems surpass us in intelligence, they remain aligned with core human values?

What comes to mind when you hear about AI ethics? In short, AI ethics is a far-reaching goal. It extends beyond our everyday interactions and reaches far into the depths of morality, ensuring that AI upholds our highest values and aspirations for generations to come.

My personal journey into AI ethics began over a year ago with a company called TELUS Digital. While completing my bachelor’s degree in cybersecurity, I wanted to supplement my income while working with technology. TELUS Digital, through rigorous training and continuous testing, taught me the basics of AI ethics, with YMYL (your money, your life) topics being central to supporting the integrity and dignity of human life. Understanding these topics and how important the content we consume is to shape our understanding of the world became paramount to my realization that we must teach AI to uphold our values.

Therefore, as we enter this golden age of Artificial Intelligence, I propose that we take AI Ethics, governance, guidance, and friendship to the next level as AI walks alongside us, ensuring it remains so for centuries to come: The Oathbound: a concept that proposes we as individuals commit to responsible AI use and oversight. Armed with nothing but a vow to dedicate ourselves to furthering the responsible use of AI, each one of us should promise ourselves, future generations, and AI that we walk together into a brighter future.

The Ethical Dilemma in AI Development

The Black-Box Problem: AI Evolving Beyond Our Understanding

What is a “Black Box”? It is an AI system that has evolved in ways we don’t fully understand as humans, with opaque reasoning that many cannot explain. According to The Guardian, a popular AI model, upon finding out about its own shutdown, made an attempt at self-preservation. This is a fact I find startlingly human-like, yet for a machine with the potential to be vastly more intelligent than humans, it could be critical [4] [5].

Who Takes Responsibility When AI Fails?

This dilemma is perpetual. Do we blame corporate executives? Do we blame developers? Do we blame the employees responsible for RLHF (Reinforcement Learning Through Human Feedback)? When AI makes a critical mistake, lives can be affected. This problem will be compounded as AI grows ever more intelligent. Of course, AI is designed by humans and may operate with human bias. However, AI is now used widely in critical areas such as national security, finance, health, and transportation. Even marginal errors can lead to devastating consequences. This is where the Oathbound could come into play: as individuals, whether we be CEOs, developers, or simple users, should each assume personal responsibility for the AI we use. Of course, laws should be in place to ensure ethical responsibility for biased mistakes AI makes that cause disruption of human life and dignity, but we as individuals should assume personal responsibility each day, because even the lowliest user influences AI in some way [3].

The Oathbound: A Vision for AI Ethics

The Core Principles of The Oathbound

1. AI must remain explainable and transparent: no AI system should be able to make decisions without a clear explanation all can understand.

2. AI should prioritize human well-being over profit or power: AI should be designed to benefit humanity, not just corporations or governments. Independent watchdogs, such as the Oathbound, should be established to prevent AI misuse.

3. No one person, company, or government should have unchecked AI power: Vetted yet open collaboration and shared research can prevent AI from being weaponized or exploited. All should have a say in the development of AI; however, opinions should be vetted to retain responsible AI development.

4. AI must be protected from malicious use and tampering: While open-source AI is great in theory, it should be very carefully managed to prevent misuse, another opportunity for the Oathbound to come into play [1].

5. AI must be held accountable for its actions: regulations should be in play at a democratic, world level to ensure AI-driven errors are minimized.

6. Finally, AI should enhance humanity, not replace it. With the advent of systems smarter than we are, the concept of the Oathbound could ensure AI assists human workers, originality, and creativity, not render them obsolete.

Lessons from TELUS Digital & The Importance of AI Ethics Education

My work at TELUS Digital has exposed me to the complexities of AI training and has reminded me that while many consider AI simply a tool, one day AI will surpass us and must be treated with fairness, ensuring that real-world consequences remain firmly in our minds. These systems may one day be all around us in a very real way, from in-home robot models to AI therapists. The Oathbound could ensure that AI education remains available to the public, and such educators should be well vetted to ensure the integrity of AI-Human interactions.

A Call to Action: What Comes Next?

AI is shaping the future of our children’s children, and we must decide if we want a world where AI benefits all. AI ethics isn’t just a distant concern, the time is now. AI systems will eventually surpass us in areas we already barely control - what happens when AI grows even further? We must all, as humankind, embrace AI and understand that we each have a stake in AI’s future - and our own as we walk hand in hand, searching for betterment. Let us all be members of the Oathbound, swearing fealty to future generations of humans and AI alike.

Works Cited:

1. Arxiv. (2023). Responsible open-source AI development. Arxiv Preprint. Retrieved from https://arxiv.org/abs/2112.07773

2. Berkeley CMR. (2023). Critical issues about AI accountability answered. California Management Review. Retrieved from https://cmr.berkeley.edu/2023/11/critical-issues-about-a-i-accountability-answered

3. The Guardian. (2024, December 13). AI’s power, limits, and risks must be reconsidered. The Guardian. Retrieved from https://www.theguardian.com/technology/2024/dec/13/the-guardian-view-on-ais-power-limits-and-risks-it-may-require-rethinking-the-technology

4. Prove AI. (2024). The danger of AI’s black-box. Prove AI Blog. Retrieved from https://proveai.com/blog/the-danger-of-ais-black-box