Introduction

This article recounts my independent study on AI ethics and how it revealed a field grappling with complex questions about fairness, accountability, and human values. Experts often disagree on the most important issues; computer scientists focus on measurable fairness metrics, philosophers contemplate AI character and personality, and critical scholars challenge the underlying assumptions behind "progress." Mohamed et al.'s decolonial perspective made me rethink how AI development can inadvertently harm global communities. I realized that the diversity of perspectives in AI ethics is a feature, not a bug; we need both technical and critical viewpoints to ask who defines what counts as bias.

The Implementation Gap

Studying how organizations implement ethics showed that even well-intentioned efforts can fall short. Initiatives like Anthropic's Constitutional AI, OpenAI's safety protocols, and Google's model cards represent genuine attempts to address ethical concerns, yet critics like Gebru argue that true ethical AI requires independence from profit-driven structures. The "Stochastic Parrots" paper highlighted that focusing on mitigating bias in models may distract us from questioning whether certain systems should be built at all. I sympathized with ethics leaders who try to improve systems from within while understanding that meaningful change may require more radical approaches.

AI Ethics in the Real World

The case of Jerome Dewald, a cancer survivor who used an AI-generated avatar in court, demonstrated that AI ethics is not abstract. Dewald believed an AI avatar would help him present his case effectively, but the judge felt misled. The case underscored how AI ethics involves trust, transparency, and human dignity playing out in real situations.

The Good, the Challenging, and the Hopeful

My research highlighted positive examples such as Be My Eyes, Climate AI, and healthcare applications that genuinely help people. These projects:

• Kept humans in control
• Were transparent about limitations
• Were built through partnerships with affected communities

Ethical AI isn't a checklist; it requires ongoing dialogue across disciplines, integration of ethics into development processes, and a recognition that humans must remain central.

What I Learned

Key lessons from my study include:

• Ethical AI is an ongoing conversation — not a destination
• Implementation matters more than intention — good policies mean nothing without execution
• The human element is irreplaceable — technology should serve people
• Everyone is a stakeholder — AI ethics affects all of us in daily life

Looking forward, balancing innovation with caution and technical solutions with social change will require constant vigilance and diverse perspectives. The future of AI ethics will emerge from many small decisions made by people trying to build technology that serves humanity.