NSW govt flexes AI policies in wake of ChatGPT

Past efforts to address “less impressive” forms of artificial intelligence have helped the New South Wales government respond to the challenges and opportunities presented by ChatGPT, according to the state’s chief data scientist.

But the arrival of sophisticated generative AI has already seen the government modify its advice to public servants through complementary information that is expected to be updated regularly as the technology develops.

NSW chief data scientist Ian Opperman, who has been instrumental in the government’s AI policy push to date, this week said the collection of policies and frameworks developed from 2019 meant the state was better prepared than it would have otherwise been.

“When ChatGPT came along, and everyone was suddenly surprised and impressed by what we could do, we did have to remind people that in NSW we have an AI strategy, and AI ethics policy and also an AI assurance framework,” he told a joint Trans-Tasman Business Circle and Infosys webinar.

“So, what we’ve been doing is helping people remember what we’ve already put in place around the less impressive uses of AI, but also provide a little bit of guidance as to what we really must do from a data privacy, data security, cybersecurity perspective.”

Ian Oppermann
NSW government chief data scientist Ian Oppermann

NSW introduced its inaugural AI strategy in September 2020, followed by an AI assurance framework previously described by Mr Opperman as a “soft stick” to assist state government agencies to design, build and use AI.

All projects that use bespoke AI systems are also required to be assessed under the framework before they are deployed to prevent a repeat of Revenue NSW’s unlawful use of an automated system to recover unpaid debts between 2016 and 2019.

Mr Opperman said that while the government has left public servants to “recalibrate” in response to ChatGPT, it was working to embed some “complementary information” to the existing policies and frameworks.

He said the information would help the government navigate some of the risks associated using ChatGPT, such as around cybersecurity and privacy, adding that people should “not put confidential personal information into ChatGPT and expect to get a new policy for NSW”.

“It’s just really a matter of re-grounding people and providing a little bit of extra context for these amazing tools that have come out, and I imagine we’ll have to do that again when the next generation… of tools come out,” Mr Opperman said.

“But principles and policies we’ve got work well. The adaptation to take into [account] some of the new things we’ve suddenly seen amplified through these new tools is what we constantly need to be doing.”

Mr Opperman also said the NSW government has “run a number of town hall meetings” to help understand “what we should and should not do with it, and the sort of areas that we can try experimenting with”.

“Most AI systems up until now, especially those that you interact with as digital assistants or with chatbots, have just been really disappointing. These are not, these are actually pretty impressive and what they can do,” he added.

Do you know more? Contact James Riley via Email.

Leave a Comment

Related stories