AI can give you an answer, but that doesn’t mean you’ve got legal advice

There is nothing inherently wrong with that.

AI is fast, accessible, and increasingly accurate at summarising legal concepts. But problems arise when the line between information and legal advice becomes blurred.

Legal risk usually sits beyond the question being asked

Across commercial, property, and employment matters, we see a consistent pattern. Clients come to us with a position they are already comfortable with, often supported by an AI‑generated answer that appears clear and well‑reasoned.

What is less obvious is what sits outside that answer.

Legal risk rarely turns on general explanations of the law. It usually turns on the details that weren’t uploaded, mentioned, or recognised as significant – and those gaps are where AI is least reliable. An answer may be factually correct on its own terms, but without the legal context to frame it properly, it can be misinterpreted or given more weight than it deserves.

VIEW ALL

Commercial contracts: ‘standard’ doesn’t mean safe

In commercial matters, AI is often used to confirm whether a clause is “standard” or whether a contract broadly reflects market practice.

What it cannot reliably assess is whether:

• Liability is asymmetrical rather than balanced.
• Commercial risk has shifted through definitions, exclusions, or carve‑outs.
• A clause works in isolation but creates exposure once read with the rest of the agreement.

AI can summarise drafting. It cannot judge whether that drafting aligns with a party’s commercial objectives, bargaining power, or appetite for risk.

Those distinctions often only matter once the relationship deteriorates – at which point they matter a great deal.

Property matters: small drafting issues, large consequences

Property transactions are particularly vulnerable to misplaced certainty.

Clients are often reassured that:

• A condition is common.
• A lease provision “usually doesn’t cause problems”.
• A form or document is widely used.

In reality, minor differences in timing, notice requirements, statutory overlays, and interaction between documents can have material consequences. In NSW property transactions especially, compliance, time frames, and procedural steps are critical.

AI does not test assumptions against the actual structure of the deal, regulatory requirements, or the client’s long‑term objectives. Those issues tend to surface only after documents are signed or deadlines have passed.

Employment law: legality is rarely the full question

In employment matters, AI answers frequently focus on whether something is technically lawful.

That framing misses most of the risk.

Employment exposure often turns on:

• process rather than outcome
• how decisions are communicated and recorded
• whether conduct forms part of a broader pattern
• how employees, regulators, or courts are likely to interpret behaviour.

AI may confirm legality while missing procedural fairness obligations, award coverage issues, or adverse action risks. In employment law, context is often more important than correctness in isolation. A technically accurate answer, stripped of the surrounding facts, can still lead a client to the wrong conclusion about what they should do next.

Confidence is not a substitute for judgement

One of AI’s most compelling features is tone.

Answers are delivered decisively, without hesitation or qualification. There is no instinctive pause to say, “this depends” or “this could create risk later”.

That confidence can crowd out the instinct to seek advice – particularly for time‑poor business owners. By the time a problem emerges, the question has shifted from prevention to damage control.

What legal advice still does that AI cannot

The distinction between AI responses and legal advice is not speed or sophistication. It is judgement and accountability.

Legal advice involves:

• Identifying issues clients don’t yet know to ask about.
• Providing the legal and commercial context needed to interpret an answer properly.
• Testing assumptions rather than accepting them.
• Assessing not only what is lawful, but what is practical and commercially sensible.
• Taking responsibility for how the law applies to a specific situation.

That responsibility is backed by professional obligations and experience with disputes when things don’t go to plan. AI does not carry those consequences.

Using AI without creating blind spots

AI is here to stay, and used properly, it can be a valuable support tool. The strongest outcomes occur where AI helps frame issues and gather information – but decisions are tested before they are acted on.

The most effective use of AI is not replacing advisers – it is knowing when an answer should prompt a deeper conversation.

That pause is where value is created:

• Before contracts are signed.
• Before decisions are communicated.
• Before positions become difficult to unwind.

In many cases, a short conversation at the right time avoids disputes, rework, and expense later.

AI can produce answers. Legal advice carries responsibility for how those answers are applied – commercially, practically, and over time.

That distinction remains critical, no matter how sophisticated the technology becomes.

Zoe Butler is a senior associate with Osborn Law.