Artificial Intelligence Usage in the Western Australian Government - 20230509001¶
Overview¶
DGov has observed the rise in demand for usage of Artificial Intelligence (AI) chatbots such as ChatGPT and Generative AI for various use cases within the WA public sector. Since AI chatbots are constantly learning from their users, any information divulged while using it is retained which creates an issue should sensitive information be transmitted.
Additionally, there is a lack of governance addressing the privacy and confidentiality issues surrounding the use of AI chatbots which allows AI organisations to collect and process data in an unregulated fashion and therefore, increases the risk of data misuse.
What Has Been Observed?¶
Currently, there are no evidence of AI based breaches affecting the Western Australian Government. However, several international corporations have had accidental information exposure which led to sensitive information being leaked. See the following news articles for reference:
- Samsung tells employees not to use AI tools like ChatGPT, citing security concerns - The Verge
- OpenAI: ChatGPT payment data leak caused by open-source bug (bleepingcomputer.com)
Recommendations¶
DGov recommends that agencies conduct a thorough risk assessment, and that information classification is considered prior to using AI chatbots. It is recommended that only information classified as UNOFFICIAL be used. Further, agencies should incorporate cyber supply chain management as part of their overall risk assessment to understand the risks surrounding the use of AI chatbots.