Online commenters criticised a National Gallery of Victoria poster for Ayoung Kim's video installation, which was made with generative AI. The critics did not separate artists using AI from AI replacing artists. Media theorist Joanna Zylinska distinguishes popular AI art, which celebrates the technology's visual spectacle, from critical AI art, which exposes its infrastructure and effects on labour. Theorist Victor Burgin argues that critical art needs a backstory and should resist stereotypes. The writer, a creative researcher, says artistic experimentation may reveal the limits of AI.
Many Australians now use generative AI tools such as ChatGPT and Gemini to find deals, pick destinations and build itineraries, with one survey putting usage at 28.8 per cent of adult travellers. A new study of 421 Australian travellers found that trust depends largely on whether AI appears reliable and whether its providers can be held accountable when errors occur. Surprisingly, greater perceived transparency was linked to a lower willingness to hand over decisions, a pattern the researchers call the transparency paradox. They suggest sharing essential information first and letting users request deeper reasoning.
Google is adding labels to YouTube Music that show when a song was created or significantly altered with generative AI. The label appears on the player page, and tapping it opens a menu with more information. Google says it is working with partners and using its own detection tools to identify AI music. Publishers can also specify, when uploading, whether their content is fully, partially or not made with generative AI, though this disclosure is optional. Google says it is not banning AI music, noting that the tools open new ways for artists to create and share music.
AU Now app
Your suburb’s news, alerts and Shorts — in one app
Alerts, Shorts and your saved stories — free on Google Play.
A new study by researchers at the University of Wollongong and Western Sydney University finds AI models now perform far better on university law exams than they did in 2023. In criminal law, AI papers averaged 76.3 per cent, beating 82.5 per cent of students. Seven of the 18 AI papers ranked in the top tenth of students. The researchers say the results raise concerns about unsupervised assessments. They propose in-person exams, assessments that require collaboration with AI, and a relay model that includes supervised revision.
Staff at three of the Big Five US book publishers say HarperCollins, Simon & Schuster and Hachette have been using AI tools without public disclosure or author consent. Workers say the tools are used to email literary agents, write publicity and back-cover copy, design cover art and create marketing materials. Several employees say publishers are also using AI to write rejection letters, which agents suspect. Some staff say they feel pressured to use AI after layoffs left teams with heavier workloads.
