Late work, academic integrity, how predictive and AI tools may be used in this course specifically, and where to get help.
Policies and support#
Late work and extensions#
Work submitted after its deadline loses 5% of the available mark per day, to a maximum of five days, after which it is marked as not submitted. Extensions are granted for documented illness or genuine emergency, and are requested before the deadline wherever that’s physically possible — a request made after the fact is judged more strictly, not because late requests are assumed dishonest, but because the course can’t verify after-the-fact claims as well as ones made in advance.
There is no penalty-free grace period. If a deadline is going to be a problem, say so before it passes.
Academic integrity#
Work you submit must be yours, with any assistance declared. That includes assistance from another student, from a predictive tool, and from a person you paid — the rule is disclosure, not prohibition, except where a specific assessment’s brief says otherwise.
Undeclared use of another person’s or another system’s work as your own is treated as a serious integrity matter, referred through the university’s standard process, independent of this course’s own marking.
Predictive and AI tools in this course specifically#
This is a course about predictive text and generative systems, so pretending they don’t exist in your own work would be a strange position for it to take. The rule is the same disclosure principle as any other assistance, made specific to what this course actually asks of you:
- In Assignment 1’s build, using a code-completion tool is expected, not merely permitted — week 5’s lecture is partly about how these tools work. State which tool you used in your submission notes.
- In audits and the final project, the audit itself must be your own analysis. You may use a predictive or generative tool to help draft prose, provided the underlying observations, evidence, and argument are yours — a generated audit that invents findings you didn’t actually gather fails the assignment regardless of how well it reads, because the spec asks for evidence you recorded, not evidence that sounds plausible.
- In week 10’s co-writing exchange, using the tool is the point of the exercise. The log you keep of what you accepted and rejected is the submitted evidence, not something to tidy up afterwards.
When in doubt, name the tool and describe what you used it for in a short note attached to your submission. A declared, well-reasoned use of a predictive tool is never penalised on its own; an undeclared one that turns out to matter is treated under academic integrity, above.
Getting help#
Consultation arrangements for each teaching-team member are listed on their people page. Questions about a specific week’s material are usually answered faster by the person who taught it than by the convenor, and questions about the shape of the course as a whole are the convenor’s to answer. If you’re not sure who to ask, ask anyone — they’ll point you the rest of the way.