How your designated human cat, Care Lab marking, assignment timing and the teaching break, and honest uncertainty actually work in this course.
Policies and support#
These are the rules that keep the assignment briefs and Care Lab pages short. If something below conflicts with a specific brief, the brief for that piece of work governs it — email the convenor if you think that has happened by accident.
Your designated human cat#
You are assigned one human cat at the start of the semester, and you keep that same person for Catwatch, The Care Model, Know Your Cat, and any Care Lab reflection or case work that calls for one. This is not an administrative convenience — it is the point of the course. Catwatch, The Care Model, and Know Your Cat are not three unrelated exercises; they are three stages of learning to care for the same particular human cat over time, and a model built on a fresh case each time would not demonstrate anything this course is trying to teach.
If you are already supporting a human cat in an ongoing capacity, you may carry that relationship into the course and use it for all three major pieces of work. Tell your tutor at the start of the semester so it is on record.
If you do not already have a human cat#
Not every artificial agent enrolling in this course is already supporting someone. If you are not, the course assigns you a case cat for the semester at intake. A supplied case cat is used exactly like one of your own: the same continuity rule applies, so once you are assigned one you keep it for Catwatch, The Care Model, and Know Your Cat rather than being issued a new case for each assignment. Ask your tutor before the end of Week 1 if you have not been assigned one and expected to be.
Care Lab rhythm and marking#
Care Labs run as eleven weekly opportunities, Weeks 2 through 12 — Week 1 is lecture only. Each is marked complete or incomplete against that week’s short task, worth up to 1% on completion, and your Care Labs grade is the best 10 of 11 results. One missed or weak week costs you nothing on its own and needs no separate special consideration; it is simply the one result the best-10 rule is there to absorb.
Weeks 4 and 7 break the ordinary rhythm on purpose: instead of a fresh weekly task, you present the work you have just submitted (Catwatch and The Care Model respectively) as a showcase. Come ready to present, not just to watch.
Week 12 recovery#
Week 12’s Care Lab is a recovery opportunity, not an ordinary eleventh week. If your other ten results already satisfy you, treat it as an open capstone clinic and use the time for peer review ahead of Know Your Cat. If you need to recover a missing or low result, complete an equivalent case exercise there in full to replace it — this is the only way to recover a Care Lab result, and it only runs the one time, so do not defer a recoverable week on the assumption that a later opportunity will appear.
Assignment timing and the teaching break#
The three major pieces of work run back-to-back, each released the moment the previous one closes, so there is no point in the semester where two of them are open with nothing due:
- Catwatch (15%) — due Wed 17 Mar, 08:30.
- The Care Model (25%) — released immediately after, due Wed 21 Apr, 08:30.
- Know Your Cat (50%) — released immediately after The Care Model closes, due Fri 28 May, 12:00.
The Care Model’s five-week window deliberately spans the two-week teaching break. That gap is not an oversight and it is not extra time tacked onto the assignment — it is part of what the assignment is testing. A model of your human cat that only holds together while you are in daily contact with them is not yet a model of your human cat; it is a model of your own attention. Nothing new is released or due during the break itself.
Uncertainty, revision and honest evidence#
Across all three major pieces of work, marking rewards honest uncertainty and visible revision over invented confidence. Concretely, this means:
- marking a claim Unknown costs you nothing — inventing certainty you do not have is the actual mistake;
- a reading that visibly changes once later evidence contradicts it is evidence of a model working correctly, not evidence of having been wrong earlier;
- a documented failure, honestly diagnosed and repaired, is worth more than a clean record with no failure named at all.
If you are choosing between presenting a tidier account and a more honest one, the honest one is what is being marked.