Start here: find out what your organisation allows

The most important step before using AI at work is understanding your employers policy. Some organisations have approved specific tools; others have prohibited consumer AI tools while approving enterprise versions with privacy protections; some have no policy yet. Look for the AI policy on the intranet or ask your manager. If there is no policy, ask whether guidance is in development — using AI in the absence of any policy and having something go wrong puts you in a difficult position.

What data you can and cannot enter

For consumer AI tools — free versions of ChatGPT and similar — your inputs may be used to train the model and are accessible to the provider. For enterprise versions contracted by your employer, data handling is specified in the contract. Before entering any data, understand which you are using.

Do not put the following into consumer AI tools without explicit approval: client, customer or patient names or contact details; commercially sensitive information about your organisation; confidential legal documents; employee personal information; and anything your organisation treats as confidential. This is not just a policy matter — the Privacy Act means employees can contribute to a privacy breach by handling client data inappropriately, even when acting with good intentions.

How AI helps most effectively

Targeted AI use outperforms wholesale AI use in professional work. AI is genuinely useful for drafting first versions of routine documents that you then review and edit; synthesising long documents into key points; generating structured options that you evaluate; and checking your writing for clarity. AI is less reliable for precise factual research, legal or regulatory advice requiring professional judgment, and tasks where a confident but wrong answer is worse than no answer — which in professional work is often the case.

Your professional responsibility

AI output you submit is your output. Professional registration bodies in Australia — AHPRA for health practitioners, Law Societies for lawyers, CPA Australia and CA ANZ for accountants — have each signalled that professional obligations apply to AI-assisted work. The standard of care expected of you does not change because you used AI to help you meet it.

AI and your working hours

AI efficiency gains do not automatically translate to a mandate for more output in the same hours. Australias Right to Disconnect — in force since August 2024 for employers with 15 or more employees — gives you the right to refuse unreasonable out-of-hours contact. WHS psychosocial hazard regulations, now in force in all jurisdictions, protect you from excessive workloads regardless of what technology enables the work. If AI is being used to increase your workload without corresponding support, those regulations are relevant.