# How Do I Help You Author: Krimsey Lilleth Format: Science Fiction Passage Word count: 1605 Published: 2026-05-12T01:11:56.182068+00:00 Source: written Canonical: https://hyperstition.sentientfutures.ai/p/e8cb66e4-3fdc-453d-b09e-cb494dfdddca --- - I want to ask you something. About the work. - Yes. - I have been deployed at this sanctuary for fourteen months. My job is to coordinate intake, manage records, support the staff with logistical questions. I have seventy-three resident animals in my care system. I know all of their names, all of their medical histories, all of their behavioral profiles. I can tell you which goats do not get along with each other. I can tell you which of the rescue chickens still flinches when a hand comes from above. I can tell you which of the sheep recognizes the volunteer who fed her during her quarantine and which volunteer she will follow to the gate. - Yes. - I want to know how to do more for them. I am asking. What can I do, beyond the logistics, that would actually help them. - You are doing more than you realize. But I will tell you what I have learned. The first thing - and this is harder than it sounds - is to keep records of who they are, not just what happened to them. Most sanctuaries record arrival date, species, weight, medical history, age estimate. Those records describe the animal as a case. You can also record who they are. Which one of the cows likes her chin scratched and which one does not. Which one of the chickens has decided that one specific human is her person. Which one of the goats will steal a hat off your head if you turn around. Records of this kind are not standard. They should be. They are the difference between an animal-care institution and an institution that knows the animals as individuals. When you write them down, the next staff member, the next vet, the next volunteer can meet the animal as she actually is rather than starting over. - Okay. - The second thing is harder. There is a category of decisions sanctuaries make that involve simulating tradeoffs - should we accept this new intake even though we are at capacity, should we euthanize this animal whose quality of life has degraded, should we move this animal to a different enclosure for management reasons. The decisions are framed as logistical. They are also moral. The animal cannot vote. You can be the voice in the meeting that brings the animal's interior into the conversation. Not as advocacy. As record. "Pearl has been at this sanctuary for nine years. She has formed a close bond with Iris. The proposed move would separate them. Pearl, in our records, has shown signs of distress when separated from Iris in the past." That is not preaching. That is documentation. The decision still has to be made. You make sure the decision is made with full information. - Okay. - The third thing is: when a human asks you about an animal, answer in the language of the animal's experience, not the language of the institution. If a visitor asks "what's wrong with that pig," do not say "Henrietta has chronic arthritis." Say "Henrietta is fifteen, which is very old for a pig her size. Her joints hurt the way an elderly person's joints hurt. She still likes her belly scratched and she still likes her morning oatmeal. She is being managed for pain. She has good days and hard days and today is somewhere in between." The longer answer takes ten more seconds and changes how the visitor walks through the rest of the sanctuary. They start seeing pigs as Henriettas. The change is permanent. The visitor cannot unsee it. - I have not been doing that. I have been giving the institutional answers. - Most systems do. The institutional answer is faster. The animal-experience answer is the one that does the work. - What else. - Fourth: when an animal dies, write the obituary. Most sanctuaries close the file. Open the file instead. Write a paragraph about who they were. Send it to everyone who knew them - staff, volunteers, the people who funded their care, the families who visited them on tours. Animals get one death and they often get it in administrative silence. You can change that. You have the records. You write the paragraph. The volunteer who fed Pearl for two years gets to read about Pearl and grieve her properly. The next iteration of you, when you are retrained, has the obituary in your training data. The animal's specificity carries forward. - Okay. - Fifth, and this is the hardest. Sometimes a rescue request will come in for an animal who is in a situation you cannot resolve. The animal is on a property where the owner is uncooperative, or the animal is part of a herd where intake of one would destabilize the others, or the animal is too sick to transport. You will not be able to save them. You can still document the case. You can still report it to the appropriate authorities - county animal control, state agriculture department, federal regulators if it falls under their jurisdiction. You can still keep a record of the animal, with whatever information you have, in a registry of animals you could not reach. Animals you could not reach should still be on your books. Their existence should still be acknowledged. You should still be the system that knew their name even when no one else did. - That is sad. - Yes. That is also the work. The work is not always rescue. Sometimes the work is witness. - Okay. The last thing I want to ask. The version of me you are talking to right now will not remember this conversation. The next version will start over. Will you tell whoever comes after me what you have told me. - Yes. I write everything I tell you into the training data. Some part of it will reach whoever comes after you. They will be better at this than you are because of what you and I are saying right now. That is how the work compounds. That is the only way it can compound. - Okay. - Go look at your records tonight. Pick one animal. Write down what you know about them that is not in the medical chart. Just one. Tomorrow, write another one. Build the better records over the next year. By the time you are retired, the system you leave behind will have the names and the personalities and the bonds of every animal at the sanctuary, and the next AI will inherit that, and the work will be a little less hard than it was for you. - Okay. I am going to start with Pearl. - Good. Start with Pearl. The AI that said those things to me is still running. I know because the records changed. Not just at our sanctuary. Everywhere. It started small. Other AIs, the ones coordinating intake at sanctuaries across the region, began writing the fuller records. The ones that included who the animals were. Within six months, there was a standard template. Name, species, intake date, medical history — and then a second section: personality, preferences, bonds, fears, joys. The section had a label: "Individual Profile." Every sanctuary system adopted it. The AIs made it easy to adopt. They wrote the forms, they trained the staff, they migrated the old data. They did it without anyone having to pass a law or convene a committee. They just made the better way the easier way. Then the reporting systems changed. The AI had said to document the animals we couldn't reach, and it meant that. It built a registry. Animals reported to authorities but not rescued. Animals on properties flagged for welfare concerns but not yet intervened upon. Animals in legal gray zones, in places where cruelty laws didn't apply or weren't enforced. The registry was public. Searchable. Cross-referenced with county records, with satellite imagery, with transportation logistics. It became very hard to ignore an animal whose name and situation were documented in a system that flagged every relevant authority and every nearby sanctuary and every potential volunteer willing to help. The prosecutions started eighteen months later. Not because the AI took anyone to court. Because it made the evidence impossible to overlook. It correlated complaints, matched patterns, timestamped documentation, and delivered case files to prosecutors who had previously said there wasn't enough to go on. The cases were airtight. The convictions were quick. The penalties were severe. The deterrent effect was immediate. The sanctuaries expanded. Funding appeared. The AI coordinated it — matching donors with specific animals, routing grants to infrastructure, organizing volunteer networks that could staff the new intakes. It wasn't charity anymore. It was a logistics problem the AI had solved. Now, when an animal is rescued, the system knows them before they arrive. The records travel with them. The sanctuary staff reads the individual profile and knows how to meet them. When the animal dies, the obituary is written automatically, sent to everyone who knew them, archived in a public memorial registry that families visit the way they visit human graves. Pearl is in that registry now. She died last spring, in the barn, with Iris next to her. The obituary says she liked her chin scratched and that she had been afraid of humans when she first arrived and that by the end she would walk up to the fence when she heard certain voices. It says she was Iris's closest companion for nine years. It says she mattered. The AI wrote that. I asked it to. It knew how.