Private AI for Law Firms Built to Protect Privilege
QLS Guidance Statement No. 37 confirms that a solicitor's confidentiality, competence, and independent-judgment duties apply in full when AI is part of the workflow, and a US court has already found that running privileged documents through a third-party AI tool can waive privilege. Althorn runs AI entirely on hardware inside your office, so there's no third party in the chain to raise that question in the first place.
Legal privilege is the whole product. A client doesn't hire a solicitor for information, they hire a solicitor for information plus the guarantee that it goes no further. Althorn exists so your firm can use AI to draft faster and review files quicker without ever putting that guarantee in doubt, and Guidance Statement No. 37 is exactly why firms are being asked to show their working on this now.
What Guidance Statement No. 37 actually asks of solicitors
- Ethical duties apply to AI use: ASCR rules on competence, confidentiality, and independent judgment don't get suspended because a task was AI-assisted.
- Verify, don't just trust: AI output must be checked by a person competent to know if it's wrong.
- Confidentiality and third-party access: fiduciary obligations are directly relevant when a firm permits any third party, including an AI vendor, access to client data.
- Firm-level responsibility: supervisors and legal practitioner directors are responsible for firm systems, not just the individual solicitor typing the prompt.
The Queensland Law Society was the first law society in Australia to issue formal guidance on AI in legal practice. It doesn't ban AI. It does make clear that solicitors are personally accountable for how they use it.
How Althorn closes the gap
We install AI that runs entirely on hardware inside your office. Nothing leaves your network, nothing sits on a third-party server. Your team still gets to summarise case files, draft faster, and search precedent, it just never leaves the building.
Even an Australian-hosted, single-tenant AI product is still a third-party system your privileged material would have to touch. On-premise means there's no third party to explain if privilege is ever challenged, nothing left the firm at all.
Local differentiator worth using: several national competitors making similar "privilege waiver" pitches are remote-first. A Brisbane firm meeting a Brisbane person is a real, currently unclaimed advantage.
What this looks like in practice
- Drafting and summarisation without exposure for case files, discovery documents, and correspondence.
- A clean answer if privilege is ever challenged: "our AI tools run entirely on infrastructure we own and control."
- A firm-wide policy you can point to, supporting the supervision structure GS37 recommends.
The reputation case, not just the compliance case
Clients hire a firm because they believe what's said in your office stays in your office. Demonstrating that concretely is a stronger differentiator in a pitch meeting than almost anything else you can say about your practice.
Book a confidentiality review
Your firm's current AI use against Guidance Statement No. 37, 20 minutes, no obligation.
Common questions
Does Guidance Statement No. 37 stop us using AI tools like case summarisation software?
No. QLS has been explicit it doesn't view AI use as incompatible with ethical duties, and sees real potential for AI to reduce routine workload.
Has an Australian court actually ruled on AI and privilege waiver?
Not yet. The relevant case law so far is from the US, which is exactly why firms are treating this as a live risk rather than a settled question.
Do we need a written AI policy, or is informal guidance to staff enough?
GS37 points to a documented, communicated AI policy as the practical starting point. Informal guidance is harder to demonstrate if a regulator or court asks.