Ten rules for using artificial intelligence technology have been launched in Australia but none may police what could be the biggest AI playground of all: your pocket.
The technology will be introduced to a larger audience next week when Apple adds generative AI to its smartphones, tablets and computers for the first time.
The iPhone maker will become the third major manufacturer to use generative AI as a selling point for its smartphones, promising it will supercharge its voice assistant Siri with help from ChatGPT.
But experts warn the trend, spearheaded by Samsung and Google, is coming at a time when many people do not understand the technology and its limits, and without rules for clear disclosures when messages or images have been digitally created.
Apple executives will unveil new products at the Steve Jobs Theatre in Cupertino in the early hours of Tuesday morning, including its first iPhones with generative AI installed.
The company has created its own platform for the technology, called Apple Intelligence, and promises it will be able to rewrite text messages for users, generate images, prioritise phone alerts, transcribe audio, and even create new emoji, dubbed Genmoji.
Apple has also partnered with OpenAI to extend AI tools on its devices.
Questions that prove too hard for voice assistant Siri to answer, for example, will be passed to ChatGPT if users give the phone permission to do so.
Apple is the last of the big three smartphone makers to bet on the technology after Samsung introduced Galaxy AI to its devices in January, and Google launched its Pixel 9 phone range under the banner, "Oh hi, Gemini".
Kantar global consumer insights director Jack Hamlin says generative AI is becoming the latest must-have feature in smartphones and buyers will expect AI tools to be installed in their next handset.
"Smartphone manufacturers are solidifying their AI strategies, embedding features that enhance everyday convenience and utility, aiming to make their devices indispensable," he said.
"The key challenge remains whether these features will be compelling enough to transition from being mere conveniences to essential tools."
While controversial, Telsyte managing director Foad Fadaghi says Apple's decision to partner with OpenAI is likely to prove popular with users who are already familiar with its platform.
"ChatGPT is by far the most used AI chatbot in Australia – more than 60 per cent of users choose it despite some very strong alternative offerings from Meta and X," he said.
"It will tick the boxes with regards to familiarity and possibly ease of use."
AI image tools will prove the most popular addition, he says, both for capturing better photographs and creating scenes based on text descriptions.
But Mr Fadaghi says concerns were already emerging about content created using mobile AI technology and whether recipients would be able to identify its legitimacy.
"There are no clear solutions yet for determining what's real or what's not," he said.
"Ironically, part of the solution might be more AI – it will definitely be in the commercial interests of all the manufacturers to have AI integrated into their handsets, not just to generate content but to detect AI use."
The questions arise just days after the federal government unveiled 10 voluntary guidelines for the use of AI, following discussions by an expert advisory group.
The rules, which have been designed for businesses, include disclosing when AI has been used to make decisions or create content, and introducing ways for people to contest AI outcomes.
The government will hold a four-week public consultation on mandatory restrictions that would apply to high-risk uses of the technology.
University of the Sunshine Coast computer science program co-ordinator Dr Erica Mealy says voluntary guidelines are unlikely to impact AI phone software, and fall short of establishing clear disclosures for the technology.
"There should be some regulation on this and there should be watermarking on imagery," she told AAP.
"The ability for the layperson to tell what is AI-generated and what is fact is getting a lot harder."
Integrating generative AI tools into smartphones will introduce more people to the technology, Dr Mealy says, which could deliver both benefits and drawbacks.
On the downside, users could take advantage of the technology to compose reports they were meant to write themselves or create deepfake images to ridicule others.
On the upside, she says, writing and translation tools could assist people with disabilities or people who use English as a second language.
"The all-AI-is-bad view is a very privileged, ableist kind of viewpoint because it does have an ability to help in some circumstances, but I do worry about younger people, older people and people with disabilities who don't have that critical lens to be able to look at it and see if it's leading them astray," Dr Mealy said.
"Putting AI in phones just makes it so much more accessible to everybody and we need more education. It's about digital literacy, not just regulation."
