The Mirror We Hold Up to AI: Ethics Starts With Us

By Chris Daily

We are living through one of the most consequential technological shifts in human history. Artificial intelligence is no longer a science fiction premise. It's writing our code, answering our questions, and increasingly shaping how we understand the world. And as AI becomes more embedded in our daily lives, a critical question demands our attention: Who is responsible for making it ethical?

As a human-centered AI evangelist, I've thought a lot about this. My answer might surprise you: the most important ethical actor in the AI story isn't a company, a philosopher, or a regulator. It's you.

The Case for Corporate Guardrails

Let me be clear — I do believe AI companies have a genuine responsibility to build systems that reflect broadly shared cultural ethics. When a company like Anthropic employs a resident philosopher to help guide how its AI responds to morally complex questions, that's not a bad idea in principle. It's actually encouraging. The alternative — an AI with no ethical scaffolding at all — is far more dangerous.

Cultural ethics represent the accumulated wisdom of communities. Ideas like fairness, harm avoidance, respect for human dignity, and honesty aren't just abstract ideals. They are the social contracts that allow us to live and work together. When AI companies bake these principles into their systems, they are, at their best, codifying a kind of shared floor — a baseline of behavior that protects users and society at large.

These guardrails matter. An AI that helps someone cause harm, spreads dangerous misinformation, or systematically advantages one group over another isn't neutral — it's actively harmful. So yes, build the guardrails. Set the standards. Reflect the broad ethical consensus of the cultures these tools serve.

But Here's the Problem

No company — however well-intentioned — can fully represent the moral universe of every person who uses their product. Ethics is not a software specification. It is not a 30,000-word document, however thoughtfully written. And it is certainly not the exclusive intellectual property of any one philosophical tradition or any one person's worldview.

When a single organization decides what is "right" for millions of users, something quietly dangerous happens: the immense diversity of human moral experience gets flattened. Your values, shaped by your culture, your faith, your family, your lived experience — they may not be represented in that document. And you may never know.

There's another risk, too. When we outsource our ethical thinking to an AI system, we can slip into a kind of moral passivity. We start to ask what does the AI think is right? instead of what do I believe? That is a subtle but profound erosion of something essential to what it means to be human.

The Only Ethics You Can Fully Control

Here is the truth I keep coming back to: all we can really control is ourselves.

You cannot personally audit the training data of a large language model. You cannot force a tech company to share your values. You cannot guarantee that any AI system will behave with perfect integrity in every situation. These things are largely outside your sphere of influence.

What is inside your sphere of influence is how you show up — in your interactions with AI, with technology, and with each other.

Ask yourself: Am I using these tools in ways that reflect my values? Am I being honest in how I represent AI-generated content? Am I thinking critically about the answers I receive, or accepting them uncritically? Am I using AI to amplify my best impulses — creativity, curiosity, compassion — or my worst ones?

These are the real questions of AI ethics for most of us. Not the grand philosophical debates in Silicon Valley boardrooms, but the small, daily choices we make as users, creators, and citizens.

A Human-Centered Vision

I believe in AI. I believe it can be a genuine force for human flourishing — expanding access to knowledge, democratizing creativity, and helping us solve problems that have long seemed intractable. But I am a human-centered AI evangelist, and that word — human — carries weight.

Human-centered means that AI exists to serve people, not the other way around. It means that the design of these systems must be grounded in human values, human dignity, and human agency. And it means that no matter how sophisticated these tools become, the moral center of gravity must remain with us.

So yes — hold AI companies accountable. Demand that they build systems that reflect broad cultural ethics and protect users from harm. Advocate for transparency, fairness, and responsibility in AI development. These are worthy and important causes.

But don't stop there. Look inward. Reflect on your own values. Use these powerful tools with intention, integrity, and care.

Because in the end, the ethical quality of an AI-powered world will be a reflection of us. And that means the most important work — the work of becoming a thoughtful, grounded, values-driven person — is work only you can do.

---

Bringing your own judgment to the tools

The questions this piece raises — am I thinking critically, or accepting AI's answers uncritically? — are exactly what Think Critically with AI is built to develop.

See the course →