Social media giant Meta's confession in a survey has sparked attention: in order to train its cutting-edge artificial intelligence models, the company voluntarily gave up the use of all publicly available images, texts, and other data of Australian adults posted on Facebook and Instagram platforms. This move stands in stark contrast to the stringent AI-related laws in the European Union, highlighting Australia's continued use of the perceived "outdated" 2020 privacy law framework, thereby inadvertently increasing the risk of Australian citizens' data being extracted and utilized by technology companies like Meta without consent.
During the in-depth investigation, Meta's Global Privacy Director, Melinda Claybrooke, initially denied the use of Facebook and Instagram user post data to train AI models. However, under continued questioning from lawmakers, she eventually had to admit to this fact. Notably, Labor Senator Tony Sheldon initially asked whether it involved Australian posts since 2007, to which Claybrooke denied; but then Green Senator David Shoebridge accurately questioned and pointed out that Meta actually downloaded all publicly available posts since 2007 that were not set as private, leaving Claybrooke with no choice but to confirm. It is worth noting that Meta's actions did not touch the data of users under 18 years old, unless these teenagers' pictures were posted on adult accounts.
Meta has a long history of privacy infringement controversies. In June of this year, several industry professionals, including Pulitzer Prize winner Daniel Etter, jointly wrote a letter calling for social media giants to be prevented from using real war and conflict images to train AI. Previously, Meta boasted about having a massive collection of "public" photos, claiming it as its advantage in the generative AI field.
When comparing with EU policies, Claybrooke explained that Meta's provision of data usage opt-outs in the European Union was entirely due to strict local laws, while in Australia, such options do not exist because the law does not require it. She further revealed that due to the strictness and uncertainty of EU laws, Meta even suspended the introduction of some AI products. Despite this, Australian users can still protect themselves by setting their data as private, but this measure comes at a time when the federal government is considering banning children from using social media to reduce harm.
Green Senator Shoebridge directly pointed out the government's inaction in privacy protection legislation, criticizing its failure to effectively prevent companies like Meta from abusing children's photos and videos. He emphasized that Europe is able to effectively protect citizen privacy precisely because of its strict privacy laws and called on the Australian government to follow suit and revise the outdated Privacy Act as soon as possible.
It is worth noting that the Australian government has signaled its intention to amend the Privacy Act, which is expected to be adjusted based on the review results from 2020 in order to better adapt to the development needs of the digital age and strengthen the protection of citizens' privacy rights.
