Match the task to the technology
Course overview · 4 min reading + 12 min practice, estimated
Principles and method
Language models generate outputs from patterns in their training and supplied context. Fluent text is not evidence that a claim is true. Choose bounded tasks with inspectable inputs and outputs, such as drafting a role summary from an approved brief. Avoid assuming a tool can infer candidate intentions or assess an entire person from a profile. Decide whether a simpler template or rule would solve the problem more reliably. Define the action the output will support and the consequences of an error before choosing the model or interface.
Worked example
A team asks AI to draft an interview preparation note from confirmed role facts. It does not ask the model to decide who deserves an interview or invent missing compensation information.
Put it into practice
Compare a template, a rule and an AI draft for one fictional recruitment task.
Use fictional information and keep your work in your own notes.
Compare your approach: self-review guidance
Explain which part requires flexible language and which part needs deterministic handling. A good task boundary names prohibited inferences and the human decision that remains outside the tool.
Sources and further reading
Original Academy teaching and fictional examples. These references provide context, not endorsement. Edition 2026.09; updated 2026-09-24.
- GOV.UK: Responsible AI in recruitment
UK guidance on procuring and deploying recruitment AI.
- NIST: AI Risk Management Framework
Voluntary framework for organising AI risks and controls.