Chartered Accountants Australia and New Zealand (CA ANZ) has raised concerns regarding the government's proposed mandatory artificial intelligence (AI) guardrails for high-risk applications.
In its submission, CA ANZ expressed concern about the criteria for defining high-risk AI systems under the proposed guardrails.
The organisation pointed out the uncertainty surrounding whether an AI system used by accountants to analyse data would fall into this high-risk category.
"While most practitioners offering services to consumers will use systems built on or incorporating AI, we do not consider that their use cases for AI would meet the principles to designate such systems as high risk," the submission noted according to Accountants Daily.
The submission by CA ANZ calls for additional guidance to help determine the risk level of an AI system, suggesting that examples of both high-risk and low-risk AI systems would be beneficial.
The body also seeks clarification on the definition of a deployer of an AI system, which influences the application of the guardrails.
CA ANZ has actively engaged with its members and stakeholders to gather opinions on what constitutes high risk in accounting.
The submission further noted: "However, in discussions with members and stakeholders, we noted that there was not a consistent interpretation of this critical point, particularly given the question in the proposals paper asking whether all general-purpose AI systems should be considered high risk."
The feedback is claimed to be required for shaping a "clarity, enforceability, and a balanced approach to mandate AI guardrails for high-risk AI systems."
Earlier in October 2024, CA ANZ, in collaboration with the Association of Chartered Certified Accountants (ACCA), released the 'Audit and Technology Playbook: A Practitioner's Guide', which is intended to guide practitioners through the integration of technology in auditing.
"CA ANZ expresses concerns about mandatory AI guardrails " was originally created and published by The Accountant, a GlobalData owned brand.