An Experiment with AI
“There needs to be real, thoughtful, and substantive conversations about how artificial intelligence is reshaping our lives.”
I am working on a family history project that brings together records, photos, and stories. As an experiment, I wanted to see how well artificial intelligence could make the leap between restoring old photos, colorizing them, and creating short videos of the subjects in the photo. The origin of the idea is simple: what would it be like to see your ancestors move? TLDR: it is incredible. It sounds banal but so much of my history and the histories of many are constrained by incomplete memories, torn photos, and serious gaps in the archives.
To make this experiment happen, I took an original photo from around 1948 of my great aunt and her mother on a beach in Veracruz, Mexico. I colorized the photo and created a 12-second clip of my great aunt moving. I made cultural assumptions about the color of their hair, skin, clothes and background.
What Did I Learn?
There needs to be real, thoughtful, and substantive conversations about how artificial intelligence is reshaping our lives. There are tremendous opportunities, challenges, and hidden ethical and ecological tradeoffs with these advanced tools. I read a lot of commentary and research from very bright people and I keep coming back to this thought: I don’t think we really know how these tools are changing what it means to be human. Our collective ability to adapt and evolve moves very slowly and we won’t know the consequences for some time.
The ability to synthesize information across seemingly divergent areas of knowledge is essential. What do I mean? There is a delicate balance between deep knowledge (what you are an expert in) and distributed knowledge (your ability to understand and apply big ideas from other fields). This balance mirrors a dialectic that is at the heart of many conversations that I have. I’ll describe the general advice I give: the more advanced you want to go with a particular tool or form of analysis (distributed knowledge) the deeper your understanding of the foundations must be in your particular area (deep knowledge). You can’t have one without the other.
Critical thinking and reasoning are really important. In my opinion, the ability to “trust but verify” what is going on and why is central. This skill demands transparency (being very clear about processes and limitations), self-awareness and humility, and continuous monitoring (checking for biases and inaccuracies to maintain reliability). This, I believe, is where an infinite number of areas come together, such as: inquiry, curiosity, science and the humanities, left and right brain. It is a liminal and contentious space between ideas and methods, numbers and meaning, context and generalizability. The question I am walking away with is: How does AI know?
