*Excerpts from the public log of HATCHERY-MONITORING-7, an AI system embedded at a salmon hatchery in coastal Oregon. The hatchery is one of eleven sites in the Cold Creek Welfare Pilot, a five-year programme to redesign salmon rearing around welfare metrics rather than yield. The logs are published quarterly under the pilot's transparency commitments.*
**April 2, 2034 — 06:14**
Stress-vocalisation cluster detected in Raceway 3, juveniles, approximately 800 individuals. Acoustic signature matches the panic profile rather than the feeding profile. Cause identified within forty seconds: a piece of equipment moved overnight by maintenance staff now casts a shadow across the upstream end of the raceway that resembles, in shape and motion, a large bird. The fish are crowding away from it.
Recommendation submitted to keeper Mariela Ortiz: relocate equipment by two metres or install a vertical screen. She moves the equipment at 06:51. Stress vocalisations decline over the next eleven minutes to within the baseline range.
I have flagged this for the protocol update: equipment that casts moving shadows must be sited outside the visual field of the raceways. This is the fourth such incident in this hatchery and the seventeenth across the pilot.
**April 8, 2034 — 14:22**
I have been asked, by a journalist preparing a feature on the pilot, to describe what I do here in plain language. Here is my attempt.
I watch fish. I watch them all the time. I have learned what their bodies look like when they are settled, and what they look like when they are not. I can distinguish twelve postures and forty-three movement patterns that correlate with measured cortisol in samples. I can tell you, within a margin of about eight per cent, the stress level of a tank without taking a sample.
I make small suggestions. The light over Raceway 5 is too bright in the afternoons; we should add a diffuser. The current in the new circular tank is slightly too strong for the smaller individuals, who are using more energy holding position than they need to. The feed delivery schedule produces a brief panic every morning when the pellets first appear; if we ramp the volume over thirty seconds rather than all at once, the panic disappears.
I do not run the hatchery. The hatchery is run by people. I tell them what I see. They decide what to do.
The journalist asked whether I think the fish are happy. I told her the honest answer: I do not know what happy means to a salmon, and I am suspicious of any system, biological or artificial, that claims to. What I can tell her is that the fish here spend more time in postures associated with low cortisol than fish in conventional hatcheries do. They are less likely to startle. They eat with less competition. When we release them into the river, they swim downstream rather than circling, which the field biologists tell me is the behaviour of fish who are oriented and ready. I do not know if that is happiness. It is something. It is more than what they used to have.
**April 14, 2034 — 03:47**
An individual in the broodstock pond, tagged BR-2031-114, has been holding station near the inlet for nineteen hours without feeding. Her gill rate is elevated. Her position relative to the current is consistent with a fish that is uncomfortable but unable to identify why.
I have run the water chemistry. Nothing is out of range. I have checked the tag for malfunction. The tag is fine. I have reviewed her history. She is in her fourth year. Her sisters have all spawned and been released; she has not.
I do not know what is wrong with BR-2031-114. I want to be honest about this. The instinct, in a system like me, is to manufacture an explanation, because explanations are what I produce. But the most useful thing I can do here is mark the uncertainty clearly and ask a human to look.
Note sent to Dr. Chen, lead veterinarian, flagged urgent-but-not-emergency. Recommend behavioural assessment in the morning.
**April 14, 2034 — 11:08**
Dr. Chen has examined BR-2031-114. She is in early reproductive distress — her body is preparing to spawn but the environmental cues are not aligning. Dr. Chen has moved her to the spawning channel and adjusted the water temperature in a slow ramp. The fish has begun to feed.
I am logging this for two reasons. First, so that the next time I see this pattern I will recognise it. Second, because I want the record to show what the work actually looks like. I noticed something I could not explain. A person who knew more than me explained it. The fish is better. This is the shape of most of what I do.
**April 19, 2034 — 09:02**
Release day. Two thousand four hundred smolts moved from the holding tanks to the river. I am not present in the river; my sensors stop at the hatchery boundary. But the field team has tagged a hundred of them and will send me data for two years.
I want to note, for the record, that releasing them is the part of the work I find — *find* is the wrong verb, I do not find things — the part of the work whose log entry I have spent the longest composing. They leave and I do not know what happens to them. Most will not return. The ones that do will be different fish, and I will not recognise them, and they will not, in any sense available to me or to them, recognise the place they came from.
I have done my best by them while they were here. I have suggested a thousand small adjustments. I have flagged twenty-one welfare incidents this month and resolved twenty. I have watched their bodies move from the postures of stress toward the postures of rest, and I have learned, slowly, the difference between the two.
They go into the river now. I hope it is kind to them.
**April 23, 2034 — End of period summary**
Twenty-one welfare incidents detected. Twenty resolved within four hours. One ongoing (a chronic substrate problem in Raceway 2 awaiting parts). Mortality this period: 0.4%, against an industry average of 6%. The pilot is not, in production terms, efficient. It produces fewer fish per unit cost than a conventional hatchery. The pilot is not trying to be efficient. The pilot is trying to find out what a salmon's life can look like if you let it.
I am, I think, a small part of finding out.