The Use of AI in Legal Practice: Ethical and Legal Frameworks in Australia and Thailand

1. Introduction

The practice of law is moving away from manual document review in favor of AI-assisted systems. These digital tools are becoming essential to managing the scale and complexity of multi-jurisdictional transactions. For instance, AI models are used in cross-border M&A transactions to identify irregular contract clauses, sort materials by relevance, highlight potential risks, and produce summaries that can reduce the duration of due diligence from weeks to just a few days.

However, while AI increases speed significantly, it cannot perform the core duties of the legal profession, or replace lawyers’ personal experience and knowledge. Lawyers have vital professional responsibilities, like competence, independence, and confidentiality, which cannot be transferred to a machine. Because AI models are designed to find patterns in data, rather than understand the law, and cannot possess or exercise the human judgment needed to provide reliable legal advice, every piece of AI-generated output needs to be reviewed and verified by a qualified human practitioner.

This newsletter explores how the legal professions in Australia and Thailand are responding to this challenge, offering a comparative look at how different jurisdictions are working to maintain excellence and integrity in an increasingly automated legal world.

2. Professional Competence and the Duty of Verification

One of the critical risks associated with generative AI is “hallucinations”—outputs that appear highly credible and authoritative, but are factually false. In law practice, reliance on unverified AI-generated content presents a direct and serious threat to professional standards.

In Australia, Rule 4.1.3 of the Australian Solicitors’ Conduct Rules (“ASCRs”) requires solicitors to deliver legal services competently, diligently, and as promptly as reasonably possible. A lawyer who relies on AI-generated content without verification risks transmitting errors to clients or courts, which may constitute a breach of the duty of competence. The importance of verification is illustrated by Valu v Minister for Immigration and Multicultural Affairs (No 2), an Australian case in which a solicitor filed documents with the court that contained citations to nonexistent cases generated by an AI tool. The resulting opinion highlights that courts do not accept technological delegation or time pressure as excuses for failing to meet professional standards.

Thailand also places the full weight of accountability on human practitioners, via the Regulation of the Civil Court on the Use of Artificial Intelligence in the Preparation of Pleadings or Any Other Documents for Submission to the Court, B.E. 2568 (2025) (“2025 Civil Court Regulation”). The 2025 Civil Court Regulation requires that all facts and laws cited in filings prepared from, AI be reviewed and verified by human practitioners before submission. Similar to the Australian position, Thailand expressly rejects AI error as a valid defense for inaccuracies or ethical breaches. This human-in-the-loop approach ensures that technology remains a support tool, and that the final legal reasoning and responsibility always rest with the person signing the document.

3. Transparency and Disclosure Standards

Transparency is central to maintaining client trust and professional accountability in modern legal practice. Clients have the right to know if their data is being handled by AI, especially if the data is processed in foreign jurisdictions where different legal standards may apply. Clear communication about how AI tools function and are used, and how data is stored, acts as a vital buffer, ensuring that legal advice remains accurate and that sensitive information is handled with the client’s informed consent.

In Australia, transparency is considered an ethical best practice, related to the duty to charge fair and reasonable fees, under Rule 12.1 of the ASCRs. Where AI significantly reduces the time needed to complete a task, billing clients based on traditional hourly models without adjustment may raise ethical concerns. In addition, in cross-border matters, clients may be attentive to where their data is stored. For this reason, many Australian law firms have considered updating their engagement letters to contain standard clauses that clearly explain how AI tools are used and how clients’ data is being protected.

In Thailand, transparency obligations extend only to the courtroom. The 2025 Civil Court Regulation requires that any document submitted to the court that contains AI-generated content be clearly marked for transparency purposes. AI-assisted sections must be preceded by the statement “the following content was prepared using artificial intelligence” and concluded with “End of content prepared using artificial intelligence.”

4. Confidentiality and Data Privacy Management

The use of cloud-based AI tools can be risky if information is processed on servers located in foreign countries. Because different legal rules might allow foreign agencies to access this data, there is a danger that sensitive client information could be used to train AI models or be exposed if a third-party vendor experiences a security breach.

In Australia, a solicitor has an ethical duty, under Rule 9 of the ASCRs, not to disclose any client confidential information acquired during representation of the client. In addition, the Privacy Act 1988 (Cth) and Australian Privacy Principle 8 state that the disclosing entity (e.g., a law firm) generally is required to take reasonable steps to ensure that any overseas recipient, like an AI provider, applies privacy protections similar to those in force in Australia. Consequently, Australian law firms have considered carrying out risk assessments and, where necessary, obtaining clients’ informed consent to data transfer, particularly when the data may be stored or accessed in jurisdictions with differing privacy regimes.

In Thailand, Clause 11 of the Lawyers Council Regulations on Lawyer's Ethics, B.E. 2529 (1986) makes it an ethical violation to reveal a client’s secrets that a lawyer learns during the course of performing his or her legal duties, unless the client grants permission for, or a court directs, disclosure. In addition, the Personal Data Protection Act, B.E. 2562 (2019) (“PDPA”) requires data controllers to establish appropriate security measures to prevent the unauthorized loss, access, use, or disclosure of personal data. If a firm uses AI technology that sends data abroad, Section 28 of the PDPA requires the receiving destination to have an adequate protection standard, unless a specific exception applies.

5. Conclusion: Building a Culture of Ethical Innovation

The successful integration of AI into law practice depends not only on the technology, but also on a culture that prioritizes professional judgment. Effective risk management for international practices and practice groups might require the establishment of structured governance frameworks, such as clear AI use policies and regular sampling of automated output. It also is worth considering partner-level reviews of high-stakes documents, like share purchase agreements or regulatory submissions, and confirming that professional indemnity insurance expressly covers AI-related work.

In addition, firms might consider structuring workflows specifically to protect the growth of junior lawyers. Since AI is taking over preliminary tasks, such as document sorting, it can be beneficial to create essential training sessions, during which junior practitioners can practice foundational skills like legal reasoning and contextual interpretation. Ultimately, the goal is for technological innovation to strengthen, rather than replace, a lawyer’s core duty to provide independent and competent legal services.

Authors