Guide · Integrity
AI and academic integrity: a checklist for colleges that want AI used for learning, not for cheating.
Banning AI does not work, and allowing any chatbot means assignments stop measuring anything. There is a middle path.
The problem
General-purpose chatbots will write an essay, solve a ledger problem or draft a project report in seconds. If students use them that way, marks stop reflecting understanding, and faculty cannot tell whose work they are grading.
A checklist for the college
- Write a clear policy on which AI uses are allowed for which kind of work, and share it with students.
- Separate practice from graded work. Allow AI coaching for practice; set rules for graded submissions.
- Prefer tools that coach rather than answer for anything students use for coursework.
- Assess understanding, not just output: short oral explanations, teach-backs or viva questions on submitted work.
- Use real, local cases in assignments, which are harder to answer with a generic reply.
- Make numbers traceable: calculations shown step by step, with the tool used.
- Protect privacy: no ID numbers, OTPs or personal financial data typed into AI tools.
- Give faculty visibility of where students are stuck, so AI use informs teaching.
Choosing tools against the checklist
| Question | Good sign |
|---|---|
| Does it refuse to write graded work? | It coaches with questions and hints |
| Can faculty set integrity rules? | Rules per course or assignment |
| Are calculations shown? | Every number from a visible tool |
| Can students see what it remembers? | Memory they can view, edit and wipe |
How Teach-Back does it
Teach-Back is a personalised AI learning portal for commerce and finance students at colleges, universities and training institutes. Each student’s course, semester, language and goals produce a syllabus-aligned plan that is frozen once set, while the journey through it adapts to their pace. Lessons are built on real cases, such as a café’s GST, a shop’s day book or a bank statement. In the teach-back studio the student explains the topic and a Socratic assistant asks, hints and checks until the reasoning holds; it coaches graded work and never does it. Every number comes from a visible tool call to calculators or rate tables, not mental arithmetic. The assistant remembers what each student mixes up, in a memory the student can see, edit and wipe, and schedules short revisions on the phone. Faculty see cohort progress, a gap heatmap, assignments and integrity rules.
Questions
How should colleges handle students using AI?
Write a clear policy, separate practice from graded work, prefer tools that coach rather than answer and assess understanding through explanations as well as written output.
Can AI be used without hurting academic integrity?
Yes, if it coaches with questions instead of doing the work, shows its calculations and lets faculty set integrity rules for graded assignments.
See it on your own data. Teach-Back — Learn it, then teach it back. Book a 30-minute working session with an engineer.