AI Mistakes Australian Councils Should Avoid

AI tools offer real benefits for local government, but councils that rush into adoption without proper governance, training, and oversight are creating risks for themselves and the residents they serve. These are the most common mistakes Australian councils make with AI — and how to avoid them.

---

Mistake 1: Entering Resident Data into Public AI Tools

The free versions of ChatGPT, Claude, and similar tools are not appropriate for processing personal resident information. When you enter data into these tools, it may be used to improve the underlying model, and it is processed on servers outside Australia.

Councils handle sensitive resident data — names, addresses, financial information, health information, complaints. Entering this data into a public AI tool is a potential breach of the Privacy Act 1988 and relevant state privacy legislation.

How to avoid it: Use enterprise plans (ChatGPT Team, Microsoft Copilot for Microsoft 365) that offer appropriate data processing agreements. Develop clear guidelines for staff on what information can and cannot be entered into AI tools. Never enter personal resident information into public AI tools.

---

Mistake 2: Publishing AI-Generated Content Without Review

AI tools can produce plausible-sounding but inaccurate content. A council that publishes AI-generated information about rates, planning rules, or council services without reviewing it for accuracy risks misleading residents and creating legal liability.

This is particularly risky for planning and regulatory information, where inaccurate advice can have significant consequences for residents.

How to avoid it: Treat AI output as a first draft, not a finished product. Every piece of AI-generated content that will be published or sent to residents must be reviewed by a staff member with knowledge of the subject matter before it goes out.

---

Mistake 3: Skipping Procurement Processes

Some councils have adopted AI tools informally — staff signing up for free or low-cost tools without going through IT security review or procurement processes. This creates risks around data security, vendor lock-in, and accountability.

How to avoid it: AI tool adoption should go through the same procurement and governance processes as any other software. This includes IT security review, legal review of vendor contracts, and appropriate councillor awareness. Develop a simple AI procurement checklist that staff can use when evaluating new tools.

---

Mistake 4: Using AI to Make Planning Decisions

Planning decisions require human judgement, professional expertise, and legal accountability. A council that uses AI to generate planning recommendations without qualified planner review is creating significant legal and reputational risk.

AI can assist with the administrative work around planning decisions — summarising documents, drafting reports, analysing data — but the decision itself must be made by a qualified professional.

How to avoid it: Be clear about what AI is and is not being used for in planning processes. AI assists with documentation and analysis; qualified planners make decisions. Document this distinction in your AI usage guidelines.

---

Mistake 5: Not Telling Residents When AI Is Being Used

Residents have a reasonable expectation of knowing when they are interacting with an automated system. A council that uses AI to respond to resident enquiries without disclosing this is undermining trust.

This applies to chatbots, automated email responses, and AI-generated correspondence. Residents should be able to tell when a response has been generated or assisted by AI.

How to avoid it: Clearly label automated responses and AI-assisted communications. Develop a transparency policy that sets out how the council uses AI and how residents can find out more.

---

Mistake 6: Failing to Train Staff

AI tools are only useful if staff know how to use them effectively and responsibly. Councils that deploy AI tools without training staff risk poor-quality output, privacy breaches, and staff frustration.

How to avoid it: Invest in staff training before deploying AI tools. This does not need to be expensive — informal sessions, online resources, and peer learning can be effective. Develop simple guidelines that staff can refer to when using AI tools.

---

Mistake 7: Assuming AI Output Is Neutral

AI tools can reflect biases present in their training data. In a local government context, this could mean AI-generated content that inadvertently favours certain groups, uses language that is not inclusive, or makes assumptions that do not reflect the diversity of the community.

How to avoid it: Review AI-generated content for bias and inclusivity, particularly for community-facing communications. Be aware that AI tools are not neutral — they reflect the data they were trained on.

---

Mistake 8: Over-Relying on AI for Sensitive Communications

Some council communications require a human touch — responses to complaints, communications about sensitive planning decisions, messages to vulnerable residents. Using AI to generate these communications without significant human input risks producing responses that feel impersonal or miss important nuances.

How to avoid it: Identify the types of communications where AI assistance is appropriate and where it is not. Sensitive communications — complaints, hardship cases, complex planning matters — should be drafted primarily by humans, with AI used only for minor assistance.

---

Mistake 9: Not Keeping Records of AI Use

Councils are subject to records legislation and accountability requirements. If AI is used to produce official documents, correspondence, or reports, there should be a record of this. This is important for accountability and for responding to information requests.

How to avoid it: Develop a records management approach for AI-generated content. This might include noting in document metadata that AI was used in drafting, or maintaining a log of AI tool use for significant documents.

---

Mistake 10: Treating AI as a Cost-Cutting Tool Rather Than a Capability Tool

Councils that adopt AI primarily to reduce headcount risk undermining staff morale, losing institutional knowledge, and ultimately delivering worse services to residents. The councils getting the most value from AI are using it to free up staff time for higher-value work — not to replace staff.

How to avoid it: Frame AI adoption as a way to improve service quality and staff capability, not primarily as a cost-cutting measure. Involve staff in AI adoption decisions and be transparent about how AI tools will be used.