Table of contents
Your professional obligations, and risks, as a tax agent
A tax agent may be subject to the Code of Professional Conduct issued by the professional body of which they are a member, and disciplinary action may be taken if they fail to observe such codes.
For example, the Ethics Pronouncement (EP) 100 Code of Professional Conduct and Ethics (EP 100), issued by the institute of Singapore Chartered Accountants (ISCA) and revised on 20 March 2025, establishes ethical requirements for ISCA members. EP 100 adopts the International Ethics Standards Board for Accountants’ (IESBA) Tax Planning Standards and from 1 July 2025, it includes new sections 280 and 380 to address tax planning and related services. Under the new section 380, a professional accountant shall recommend or advise on a tax planning arrangement for a client only if the accountant has determined that there is a credible basis in laws and regulations for the arrangement.
As for tax agents who are members of the Singapore Chartered Tax Professionals (SCTP), they must comply with the Code of Professional Conduct and Ethics of the SCTP. The SCTP’s Code of Professional Conduct and Ethics contains several fundamental principles adopted from the Handbook of the Code of Ethics for Professional Accountants issued by the IESBA in 2014. One of the fundamental principles is that of “professional competence and due care”, under which “a member has a professional duty to carry out his work within the scope of his engagement and with the requisite skill and care”.
Furthermore, tax agents may also be sued by their clients under common law for breach of contract or negligence for incorrect advice, resulting in costly awards of damages.
Your professional obligations, and risks, as a lawyer
Under various legislation, rules, codes of conduct and practice directions, lawyers must ensure that the materials they put before the courts are independently verified, accurate, true and appropriate.
On 23 September 2024, the “Guide on the Use of Generative Artificial Intelligence Tools by Court Users” was published on the Singapore courts’ website. The guide sets out general principles and guidance in relation to the use of generative AI tools in court proceedings. Based on this guide, court users must ensure that any AI-generated output used in their court documents is accurate, relevant and does not infringe intellectual property rights.
If a lawyer does not follow what is stated in the guide, the courts may, amongst other things:
- order costs against the lawyer
- disregard any document or other material submitted to the courts (in part or in full), or give less evidentiary weight to the document or material in coming to a decision
- take disciplinary action against the lawyer
- take the appropriate action in accordance with existing laws in respect of intellectual property rights, personal data protection, the protection of legal privilege and contempt of court.
Lawyers may owe duties of care to their clients in both tort and contract and may be sued for negligence or breach of contract, with the awarding of damages against them.
How relying on free and low-cost AI technology puts you at risk
The role of tax and legal professionals is to help their clients navigate the complexities of regulatory compliance and legal obligations, whilst ensuring the best outcome based on their clients’ unique circumstances. At the heart of this role are 2 core functions: advising the client by recommending a course of action, outlining the potential risks and benefits; and ensuring compliance obligations are met by completing forms or tax returns on the client’s behalf.
To perform this role to the standards required by the governing legislation, you must act competently and with reasonable care. This is why it is crucial that the research materials you use as the basis of your advice or compliance activities are comprehensive, accurate and up to date so that you can trust the results. Relying on free or low-cost solutions puts you at risk of providing the wrong advice or completing compliance tasks with inaccurate information; alternatively, you or your staff have to spend additional time validating the answers from other sources to minimise the risk of providing inaccurate advice to your clients.
Out-of-date data sets
Some free or low-cost AI tools are trained on static datasets that are from a particular point in time and not updated. This limitation is often not transparent to the user. As a result, the answers they provide can be out of date and inaccurate as they are generated from a content set that does not have the most recent sources of information. This can include a missing substantial legislative update or a crucial rate or threshold change.
Wrong jurisdiction
Free or low-cost AI tools do their best to find an answer, any answer. That can include composing an answer based on information from another jurisdiction within Singapore or even from another country. On the surface, these answers appear accurate but do not reflect the correct law for your client’s situation.
Hallucinations
Generative AI is prone to hallucinations, producing an answer that reads as authoritative but which is actually totally false. In Singapore, a self-represented litigant had asked ChatGPT to give her a list of cases which were relevant to a particular legal proposition and a summary of those cases. ChatGPT provided her with 5 completely fictitious cases which looked real. However, the opposing counsel then searched for the 5 citations and found that the alleged cases in fact did not exist.
Privacy and security
There are also significant privacy and security concerns when using free or low-cost AI technology. Tax agents and legal professionals must generally seek client permission before disclosing any of their information to third parties, including AI tools. Despite training and guidelines, there is always a risk that sensitive or confidential data relating to your clients or firm may be uploaded into the AI tool.
Not all AI tools have the same level of privacy and security and some may:
- be vulnerable to unauthorised access or hacking if they do not store or transmit data securely
- acquire personal or usage data without clear consent and without being transparent about who has access and how it is used, or
- share data with third parties without explicit consent.
Once lost, such data can never be retrieved and, if client information is involved, the loss of trust may also never be retrieved.
When accuracy matters, trust isn’t optional
Adopting AI technology for your professional practice is a business investment decision that must be implemented with due consideration of the risks and benefits. Short-term cost savings might improve the immediate bottom line but could have long-lasting negative impacts on your business and reputation, and expose you to financial penalties as well.
Wolters Kluwer has a global division of expert engineers and developers dedicated to creating the best AI solution in the market for tax and legal professionals, complying with world-standard protocols around privacy and security. Your data, and your clients’ information, are safe with us. Extensive work has been done to reduce the risk of hallucinations in AI-generated answers, reinforcing the trust factor.
Already in use in North America, our AI solution will soon be deployed across our award-winning CCH iKnowConnect research platform in the Asia Pacific. This AI tool will generate results you can rely on as they are sourced from the extensive content set curated and updated by the subject matter experts at Wolters Kluwer. AI-generated answers contain links back to our content set so that you can validate the result, apply the law to your clients’ circumstances and be confident in the advice you are giving. Our content set is updated daily by our experts, capturing the latest changes to ensure that you and your clients are always up to date. This is Authoritative AI, built on decades of editorial integrity and expertise, not just on algorithms.
Read more about our Authoritative AI.
