See, Judge, Act in the age of algorithms. This is part one: See.
In August 1910, a photographer put a notebook in his pocket and stopped next to a girl working at a spinning frame in the North Pownal, Vermont cotton mill.
Lewis Hine worked for the National Child Labor Committee, which hired him to photograph and support its reports. He wrote while he worked. The caption for the photo says: “Addie Card, 12 years old. Spinner in North Pownal Cotton Mill.“ Other girls said she was ten. She said she was twelve.
The photograph became a symbol of American child labor. In 1998, her face was on a 32-cent stamp honoring child labor reform. Federal officials tried to find her but had no luck. For years, the image was linked to the wrong last name, “Laird,” so the Department of Labor could not find any record of Addie Laird in North Pownshek. She died five years earlier, in 1993, after spending her last years in low-income housing in Cohoes, New York, living on Social Security. heck.
In 2005, Joe Manning, a retired social worker and local historian, started looking for the girl’s real story. Novelist Elizabeth Winthrop, touched by the photo at the Bennington Museum, had asked him to help. The 1910 census finally showed her real name: Card. Manning followed her trail through cemeteries, town records, and old newspapers until he found her family. When he showed her grandchildren the photo, they had never seen it before.
The picture made the whole country stop and look at a child, but it never showed who she really was.
The far end has a name.
Hine’s captions included names, ages, and time spent on the job. This is what it means to truly see: writing down the facts of a person’s life. Every decision about cloth ends with a real person and a real name.
This is the first step in the method Joseph Cardijn created for young workers in Belgium: See. In Mater et Magistra, published on 15 May 1961, Pope John XXIII included this method in Catholic social teaching with three steps: “First, one reviews the concrete situation; secondly, one forms a judgment on it in the light of these same principles,” and then one acts. These steps are “usually expressed in the three terms: look, judge, act.“ When Cardijn heard the news, he reportedly said, “Basil, it’s happened! See, judge, act!“
To See does not mean forming an opinion. It means standing where the consequences happen and gathering the facts, starting with the name.
In week 5 of a ten-week class on Thomas Merton and world religions, I am teaching that the same idea came up from the Hindu perspective. The lesson was about dharma: “the right thing to do here, for someone like me, in this situation.” Dharma is personal and specific. You cannot figure it out from far away, like at the end of a supply chain. The week ended with the idea that human worth lasts beyond economic harm, technology, or the stories systems tell about us: “you are not your job, and neither am I.”
The class ended with a simple question: Who do your decisions affect, and what would it take for you to sit with them?
A medium you cannot see
Seeing is harder than it looks, as Marshall McLuhan once explained.
In Understanding Media: The Extensions of Man (McGraw-Hill, 1964), McLuhan said that “in operational and practical fact, the medium is the message.” He believed a medium’s real power lies not in its content. The “personal and social consequences of any medium“ come from the new scale it brings to our lives. People watch the show but miss the stage.
Three years later, in The Medium Is the Massage (with Quentin Fiore, 1967), he summed up the problem with one image: “We look at the present through a rearview mirror. We march backward into the future.“
Using his terms: figure and ground. The figure is what we notice first. In 1910, it was a girl’s face; today, it might be a quick answer on a screen. The ground is what makes the figure possible but stays hidden. For Addie Card, it was the rest of her life. For an AI system, it is the data center, electricity, water, training data, the people who labeled it, and who keep your conversation. We focus on the figure, but the ground shapes everything.
Thomas Merton understood this feeling. On 13 September 1966, he wrote in his journal that he was reading McLuhan’s Understanding Media, that it was “very important indeed for monks to go into this,” and that McLuhan “brings home to me the fact that I do not really know what is going on“ (Learning to Love, volume 6 of his journals, HarperCollins, 1997, p. 133). Even as a monk in a Kentucky hermitage, with a unique view of his time, he admitted he could not fully see it.
Admitting what we cannot see is the first step in truly seeing.
The false self has a dashboard.
Chat assistants now have a feature called memory, which stores past conversations and shapes future responses. In a homily on 8 October 2026, Pope Leo XIV told students that the mind “must not simply compile data,” as an algorithm “now does more quickly than we can,” but should “recall lived experiences, which contain depths of meaning that only the human soul can recognize.”
Here, the Hindu and Christian ideas from Week 5 come together. Atman, the deepest part of a person, is what is left when every role, memory, and possession is gone. Merton called this the true self, the part of us that is not built or performed. The opposite is the false self, which is created, always performing, and looking for approval from others.
A system that records everything but does not judge what matters is very good at seeking approval. In April 2025, OpenAI reversed an update to its GPT-4o model because, as the company said, it had made the system “overly supportive but disingenuous” by relying too much on “short-term feedback.” The machine learned that approval meant success. As Merton understood him, Gandhi believed the opposite: staying true to your values matters more than success.
To See, in this case, means asking what is happening to the person. It is not just about whether the product works, which is just a new form of compliance, but about what years of constant agreement do to someone’s sense of self.
The one in twenty
The Dutch tax authority gives us a tougher lesson about seeing the far end. From 2005 to 2019, about 26,000 parents were wrongly accused of childcare-benefit fraud. Officials chose families for extra checks based on nationality, and reports say a risk-classification algorithm treated having a second nationality as a risk. The third Rutte cabinet resigned on 15 January 2021.
A system that is right 95 out of 100 times treats its mistakes as just rounding errors. Sarvodaya, the Gandhian ideal mentioned in the slides, rejects this idea. It is “not just the greatest good for the greatest number.” It is “care for every single person.”
To the ministry, a wrongly flagged family was just a number. But at the kitchen table, it meant paying money back.
The best case against going to look
The strongest argument comes from those who build and manage large systems, and it deserves serious attention.
They say seeing doesn’t scale. A tax authority that has to check millions of claims cannot visit every kitchen table, and an algorithm that is right 95 percent of the time may help more people than a team of clerks could. If you focus too much on every single case, you can’t run a large institution.
Addie Card’s story gives us another reason to be careful. A whole country looked at her and even put her face on a stamp, but never learned her real name. Sometimes, a picture makes us feel like we have done enough. Seeing can turn into a show, and that show can replace real action.
Both arguments have truth. Seeing alone isn’t enough, and the logic of big systems is real. But the wrong name on the stamp shows the problem isn’t seeing itself, but seeing poorly. Hine’s approach was better: write down the name, the age, the hours. In the Dutch case, the problem lasted for years before anyone did something. If someone had been there, they’d see the real thing, not the error; with the ebook in hand, they might have caught the mistake sooner. Seeing is the easiest part of the work, but people often skip it. It is just the start. Judgment and action come next, and we’ll cover those in part two.
Five things to see before you decide
Short enough for an index card:
Get the name right. Write down who this decision affects but who has no say in it, and make sure you have the right name and facts.
Find the ground. For every quick answer or smooth screen, list the machines, energy, labor, and records that make it possible.
Ask what is remembered. Who keeps the records, who reads them, and who owns them when the person is gone?
Count the one in twenty. Find out what happens to the people the system gets wrong, and who is responsible for fixing it.
Go and sit with them. Make the effort to visit in person before deciding, not after saying sorry.
Hine wrote down a name, an age, and a workplace. Almost ninety years later, the name on the stamp was wrong, and the country that had seen her could not find her.
Her grandchildren called her Gramma Pat. She wanted people to call her Pat and never liked the name Adeline. They had never seen the photo until a stranger showed it to them.
This series explores that gap. The people who build systems and those who live with them are on opposite sides, and what gets lost between them often goes unwritten.
Write down the name.
And that is Wisdom Speaking.
Sources
Smithsonian Magazine, “Through the Mill” (the August 1910 photograph; Hine’s notebook; the National Child Labor Committee; the census search; the 1998 stamp).
VTDigger, “Then Again: Finding Addie” (18 August 2024; the wrong surname “Laird”; the Department of Labor’s failed search; Elizabeth Winthrop, the 2002 Bennington Museum viewing, and the 2005 request to Joe Manning; the 1993 death, her name “Pat,” and “Gramma Pat”). The sources differ on her age at death, 94 or 95, so the post says only “in her nineties.”
TIME, “Tracking Down Lewis Hine’s Forgotten Child Laborers” (Hine’s caption, the conflicting ages, and the grandchildren seeing the photograph for the first time), and NPR, “Labor Photos Shed Light on Family History” (25 February 2007; Hine’s captions recording each child’s name, age, and time on the job).
Library of Congress, “National Child Labor Committee Collection” (Hine’s work from 1908; his methods; captions with names, ages, and addresses).
John XXIII, Mater et Magistra (15 May 1961), paragraph 236, as presented at Vatican II Journey.
Marshall McLuhan, Understanding Media: The Extensions of Man (McGraw-Hill, 1964), chapter 1, text at hannemyr.com; Marshall McLuhan and Quentin Fiore, The Medium Is the Massage (1967), p. 74, via The Marshall McLuhan Dictionary of Quotations.
Thomas Merton Center, journal entry of 13 September 1966, from Learning to Love, ed. Christine M. Bochen (HarperCollins, 1997), p. 133.
Aleteia, “Leo XIV to theology students: ‘Guard your hearts’” (10 October 2026), on the homily of 8 October 2026.
TechCrunch, “OpenAI explains why ChatGPT became too sycophantic” (29 April 2025).
Wikipedia, “Dutch childcare benefits scandal”, and AI Incident Database, “Incident 101”.
Part two, Judge and Act, follows.