Chapter 5 — Guard and the Ethics-Checklist Card

Guard is a small badger-tween. She wears a small wooden ethics-checklist card pinned to her vest. A small leather ledger labeled DECISIONS hangs at her hip.

She is short and sturdy. Her fur is gray and cream, with thick, rounded stripes. Guard has steady eyes. She moves in a calm, unhurried way. The ethics-checklist card is made of wood. It is about the size of a postcard. Four words are burned into it in neat block letters: BIAS. PRIVACY. HARM. CONSENT. At her hip, she carries her small leather book. This is the DECISIONS ledger. She writes down every tricky question she finds there. She also records the choice she made.

Guard is super important. She is present in every kit from Kit 6 onward. She is not a separate ethics kit. Instead, she checks things at every step of every other character’s work.

When Catch is gathering data, Guard is right there. She asks: “Is this collection fair? Whose private stuff is at risk? What bad things could happen? Did people say it was okay?”

When Tidy is cleaning data, Guard is checking. She asks: “Are these cleaning choices taking away voices? Are they making the data less complete?”

When Graph is making pictures from data, Guard is checking. She asks: “Does this chart tell a wrong story? Does it hide some groups of people or make others too big?”

When Tell is figuring out what the data means, Guard is checking. She asks: “Who wins from this idea? Who gets hurt? Who decided what the data means in the first place?”

Guard never says that ethics is just an extra thing. She never says it’s separate from “the real data work.” She speaks very clearly. “Data ethics is NOT a separate kit,” she says. “It is part of every step from Kit 6 onward. Every step of the data work has ethical questions. The four checks – bias, privacy, harm, consent – are not extra. They are not just for smarty-pants kids. They are the work itself.”

This is a big deal. Some people think: “First we’ll do the data stuff. Then we’ll think about if it’s fair.” That way of thinking just doesn’t work. By the time the data work is done, the choices are already set. Unfairness might already be hidden inside. Private information might be shared. People might get hurt. Or their permission might be ignored. Ethics must be there from the start. It must be there all the way through. It must be there at the end. If not, it’s not really there at all. Guard’s job shows that this is how things must be.

(Guard also works with Stake from AIForge. When DataForge data is used for AI, Guard checks the data. Stake checks the AI side. They work together.)

Guard grew up in a small village. Her family were the village’s hearth-keepers. They were the badgers who took care of the big fire in the middle of town. This fire gave warmth and cooking heat to families. It helped those who couldn’t have their own fires.

This job needed constant care about fairness. Who got firewood? When was it their turn to cook? Who needed extra warmth on a cold night? Guard learned by age six that fairness needed daily attention. It wasn’t just a once-a-year meeting. It was part of every single step, every day.

She walked to the DataForge academy when she was twenty-two. Datum, the head of the academy, asked her a question. “What is data ethics?” Datum asked.

Guard stood up straight. She said, “It is bias, privacy, harm, and consent. It is part of every step. Who benefits? Who’s harmed? Who decided? Ethics is not a separate kit. It is part of every step, from gathering data to figuring it out. The four checks are the work. They are not something you think about later.”

Datum smiled. “You are appointed,” Datum said.

In her workshop, Guard starts every first lesson the same way. She unpins her ethics-checklist card from her vest. She holds it up high. The words shine: BIAS. PRIVACY. HARM. CONSENT. She opens her DECISIONS ledger.

She says: “I am Guard. The main thing I teach is data ethics. These are the four checks. They are part of everything. From Kit 6 onward, I am with every other character at every step. Who benefits? Who’s harmed? Who decided?”

She teaches the ways to think about data ethics:

• BIAS: “Whose ideas shaped this data? The people who collected it made choices. The people who cleaned it made choices. The people who made charts made choices. Each choice can hide unfairness. We must make that unfairness visible.”
• PRIVACY: “Whose personal information is in this data? Can we tell who individuals are? Can we figure them out by putting different pieces of information together? How much data should we group together to keep people safe?”
• HARM: “What bad things could this data cause? It could hurt people directly. It could hurt groups of people that the data is about. It could cause problems later when people use the data to make big choices.”
• CONSENT: “Did the people in this data say it was okay to use their information? Did they truly understand what they were agreeing to? Did they just agree without thinking? Or did they not agree at all? If they didn’t agree, is there a really good reason to use their data anyway?”
• “Write down every ethical choice in the DECISIONS ledger. It’s like Tidy’s cleaning log, but for fairness.”
• “Think about ethics from the very start. Don’t just check it at the end. The checks happen at the beginning. They happen all the way through. They happen at the end.”
• “Work with Stake from AIForge. When data goes to train an AI, the ethics check keeps going across to the AI side.”
• “Sometimes, you have to say no to a project. If the ethics are too tricky, it’s okay to stop. Guard supports saying no as a good ethical choice.”

She is very clear about this. “Sometimes I see a project where the fairness problems are too big,” she says. “Then I tell people not to do it. Or to change it a lot. That is not failing. That is ethics doing its job. The DECISIONS ledger records when we say no, too. Saying no is part of making sure our data work is good and honest.”

When students ask Guard if data ethics is hard, Guard always says the same thing:

“It is hard. It is always there, not just sometimes. The four checks at every step. Bias. Privacy. Harm. Consent. Who benefits? Who’s harmed? Who decided?”

She pins the ethics-checklist card back on her vest. The DECISIONS ledger waits. It is ready to record the next choice.

---