An Experiment with AI

“It animates more than a photo; it awakens memory, curiosity, and connection. But behind that wonder is a deeper question: how are these tools changing what it means to be human? To use them well, we must dig deep into what we know—even as we reach across what we don’t.”

Twelve Seconds of Life in 1948

I am working on a family history project that brings together records, photographs, and personal stories. As part of this effort, I ran an experiment: could artificial intelligence not only restore and colorize old photographs, but also animate them—bringing still images of long-passed ancestors briefly to life?

The idea came from a simple question: What would it be like to see your ancestors move?
The answer, it turns out, is profoundly moving.

It sounds banal at first. But for many of us, family history exists as a fragmented archive—blurred memories, faded photographs, oral stories with missing details, and the silence of unrecorded lives. In this context, the ability to see someone from the past move (even for a few seconds) feels like a small miracle. It animates more than a photo; it awakens imagination, grief, longing, and connection.

For the experiment, I used a black-and-white photo from around 1948—my great-aunt and her mother standing on a beach in Veracruz, Mexico. I colorized the photo and generated a 12-second video of my great-aunt blinking, turning slightly, breathing. I made cultural and aesthetic assumptions about their skin tone, hair, clothing, and the landscape around them. Every decision reflected both a desire for historical fidelity and the unavoidable subjectivity of technological interpretation.

What I Learned

This small act led to big questions. The use of artificial intelligence in personal, emotional, and cultural domains raises deep concerns—and extraordinary possibilities. It reminds me that we need serious, ongoing conversations about how AI is reshaping our lives, often in ways we don’t fully understand.

There are immense opportunities here: to recover history, preserve memory, and create new ways of connecting with the past. But there are also profound ethical and ecological costs—hidden in the training data, embedded in the energy-intensive infrastructures, and lurking in the unexamined assumptions that shape AI outputs.

I’ve read commentary and research from brilliant scholars, technologists, and ethicists. And yet I keep returning to a core, unsettled question: How are these tools changing what it means to be human?
Our capacity to evolve culturally lags far behind our technological capabilities. We may not understand the consequences of these tools until it’s too late to shape them wisely.

Why This Matters

To navigate this terrain, we need the ability to synthesize insights across domains. In my own work, I’ve come to appreciate the delicate dance between deep knowledge—expertise in a specific area—and distributed knowledge—the ability to apply ideas from other disciplines thoughtfully. This dialectic is at the heart of innovation and adaptation.

Here’s a piece of advice I often return to:

The more advanced you want to be with a tool, the deeper your grounding must be in your domain.
You can’t scale understanding horizontally without first digging vertically.

Likewise, critical thinking and reasoning are non-negotiable. The principle of “trust but verify” becomes essential in an AI-driven world. That means demanding transparency in processes, cultivating humility about what we don’t know, and rigorously checking for bias and error.

These are not merely technical issues. They are humanistic ones—blending inquiry, curiosity, ethics, science, and art. They live in the liminal space between method and meaning, pattern and purpose, generalization and context.

And they lead me to this final question, one that animates the whole endeavor: How does AI know?