Our AI guideline is on version four.
That’s not because we got the first three wrong. It’s because AI tools, and the rules around them, are changing every few months. And because our team, like yours, was already using AI before we’d written a word about it.
Microsoft’s 2026 Work Trend Index found 63% of Australian AI users are producing work they couldn’t a year ago. Yet only 28% work in organisations with a clearly aligned AI strategy and policies. That gap is where the risk sits.
Why you need a guideline now
Your people are already using AI. The question is whether they’re using it well, safely and in a way you can stand behind.
Without a guideline, you’re relying on individual judgement about:
- which tools are safe to use
- what information can go into them, and what can’t
- who checks the output before it reaches a client, candidate or employee
- when customers or staff should be told AI was used
A good guideline does two jobs. It gives people permission to use AI with confidence. And it draws clear lines around the risks.
Embrace the opportunity
The businesses getting real value from AI aren’t just speeding up old tasks. They’re redesigning work. We encourage clients to sort every task in a team into three buckets: work AI can automate, work AI can augment, and work that must stay human-centred.
Some examples of real advantage:
- For the business: first drafts of policies, procedures, proposals and reports in minutes rather than hours. Faster analysis of survey comments, exit interviews and engagement data.
- For customers: quicker turnaround, plain-English explanations of complex rules, and more time for the conversation that matters.
- For employees: less administration, help with writing and accessibility, faster answers to everyday policy questions, and more time for the human parts of the job.
Identify and address the risks
The risks are real, and most are familiar HR and legal risks wearing new clothes.
- Privacy and confidentiality: personal or client information entered into public tools may be stored or used offshore.
- Accuracy: AI can produce confident, plausible and wrong answers, including invented legal references.
- Bias and discrimination: AI used in recruitment, performance or termination decisions can embed bias, and employers remain liable.
- Transparency: 79% of Australians want to be told when AI is used, according to the OAIC’s 2026 community attitudes survey.
- Over-reliance: skills fade when people stop thinking for themselves.
The EU AI Act: a useful guide
Australian businesses can learn from Europe’s AI Act, the world’s first comprehensive AI law. It sorts AI by risk:
- Unacceptable risk, banned: including emotion recognition in the workplace.
- High risk, strictly regulated: including AI used to recruit, screen CVs, decide promotions or terminations, allocate tasks, and monitor or evaluate staff.
- Transparency risk: for example, people must be told when they’re dealing with a chatbot.
- Minimal risk: most everyday tools, with no special rules.
Obligations for high-risk employment uses have been pushed back to December 2027, but the direction is set. Penalties reach €35 million or 7% of global turnover. If you use AI in people decisions, the EU’s high-risk list is a sensible place to apply your strictest controls.
Where Australia stands
Principles. Australia’s eight AI Ethics Principles (2019) remain the foundation: wellbeing, human-centred values, fairness, privacy and security, reliability and safety, transparency and explainability, contestability and accountability.
Practical guidance. In October 2025, the National AI Centre published Guidance for AI Adoption, with six practices: decide who is accountable, understand impacts, manage risks, share essential information, test and monitor, and maintain human control. It comes with a free AI policy template.
Regulation is coming. In July 2026, the Prime Minister announced mandatory Australian Standards for AI and a new Office of AI. National Cabinet has confirmed the Commonwealth intends to legislate the standards in early 2027.
Privacy from 10 December 2026. Privacy policies must disclose when computer programs make, or substantially help make, decisions that significantly affect people. The OAIC names recruitment screening software and AI-assisted performance decisions as examples.
Existing laws already apply. The Government’s position is that Fair Work, anti-discrimination and WHS laws are technology-neutral, and employers can’t shift liability to an AI system. The Fair Work Commission now expects parties to disclose AI use in some claims.
What our guideline includes
- Why we use AI and the opportunity we see
- An approved tools list, ranked by data security
- What must never go into an AI tool, including client and employee personal information outside approved tools
- Human review: a person is accountable for every output
- When and how we tell clients we’ve used AI
- Privacy obligations, including overseas disclosure
- Training expectations for every team member
- Current legal references, reviewed regularly
Free training to get started
- Anthropic Academy: free AI Fluency courses, including versions for small businesses and nonprofits, at anthropic.skilljar.com
- Microsoft: AI Skills Navigator and free LinkedIn Learning AI courses. Microsoft has committed to train three million Australians in AI skills by the end of 2028.
- National AI Centre: Guidance for AI Adoption, with a free AI policy template, AI register template and screening tool.
- AI Adopt Program: government-funded support for SMEs.
- Infoxchange: AI learning resources for not-for-profits.
Top 5 tips
- Start with what people already do. Ask your team which tools they use and for what.
- Name approved tools. Make the safe option the easy option.
- Draw clear data lines. Spell out what can and can’t be entered into each tool.
- Keep a human accountable. Especially for decisions about people.
- Train, then review. Build AI literacy and update your guideline at least twice a year.
Frequently asked questions
Is there an AI law in Australia? Not yet a dedicated one. Existing privacy, discrimination, workplace and consumer laws apply, and mandatory AI standards are due to be legislated in early 2027.
Does the EU AI Act apply to my business? Generally only if your AI systems or their outputs are used in the EU. It’s still a useful benchmark.
Can we use AI to screen job applicants? You can, but you remain responsible for discrimination and privacy compliance. From 10 December 2026, your privacy policy may need to disclose it.
Should we ban ChatGPT? Bans tend to push use underground. A guideline with approved tools is usually safer.
How long should a guideline be? Short enough to be read. Ours is a few pages, supported by training.
How Harrisons can help
We help organisations write practical AI guidelines, map tasks for work redesign, and train leaders and teams to use AI safely. If you’d like help with yours, get in touch with our team.
Claire Harrison is the Founder and Managing Director of Harrisons, a flourishing HR consulting business that sprouted in 2009 from Claire’s passionate belief that inspiring leaders and superstar employees are the key success factor to any business. With over 20 years’ experience, Claire has worked as a HR Director of multi-national organisations, as a Non-Executive Board Director, and a small business owner. Claire’s corporate career includes working with companies such as BHP, Westpac, Fonterra and Mayne Nickless.
