AI & ChatGPT for Beginners

Learn

Before using an AI tool, classify the information. Public information is authorised for open sharing. Internal information may be restricted to the organisation. Personal information identifies an individual. Confidential or regulated information may include customer details, protected project drawings, medical records, salary data or trade secrets. The precise rules depend on organisational policy and applicable law.

Removing a person's name does not always make a dataset anonymous. A combination of job title, dates and location can reveal identity. Use synthetic examples when teaching or testing. For real data, obtain proper authorisation, check the provider's settings and contracts, and minimise what is shared. If you are unsure, do not upload the information.

Worked example

Instead of pasting a real employee performance record into a chatbot, invent a fictional case with no identifying details and ask for a neutral feedback template. Then apply the approved template in the employer's authorised environment.

Practice

Classify eight hypothetical items: a public brochure, internal project diary, client phone number, fictional expense table, published law, access password, unreleased quotation and your own practice notes. For each, write whether an external AI service may receive it and why.

Key takeaway

Protect information before you prompt, not after a disclosure occurs.