What Schools Should Ask Before Approving an AI Tool
Evaluate a tool against a specific use case. The same service may be suitable for one supervised activity and unsuitable for another. An approval conversation works better when the team can describe who will use the tool, what they will enter and what learning it should support.
Purpose and evidence
What educational need does it address? What will the learner still need to think through? Ask what evidence supports the proposed use and what the team will observe during a trial.
Learners and access
Which ages and account arrangements does the provider permit? Can all intended learners take part with the available devices and accessibility support? Identify a workable alternative for anyone unable to use the tool.
Information and settings
What information would staff or students enter? Ask the responsible team to review the provider's current terms, settings, retention and training-use explanations. Avoid entering identifiable learner information simply to test a feature.
Output and judgement
How will students check claims and recognise unsuitable or biased output? Who helps when a response is confusing or concerning? Consider the actual activity rather than assuming one control solves every problem.
Authorship and sharing
What use is permitted in assessment? How should students acknowledge assistance? Clarify the conditions for saving, exporting or publishing their work, including any separate permissions for identifiable student material.
Costs and ongoing ownership
Identify required paid features, limits, renewal arrangements and the person responsible for maintaining the approval. Record the decision, restrictions and next review date. If a relevant feature or term changes, reassess the approved use.
A practical record can fit on one page: intended activity, learner group, approved account model, information restrictions, educator checks, alternatives, owner and review date. Work within your school's existing policies and obtain the advice your setting requires.
Comments