8 hours ago
7
Meta has urged the Australian government not to make privacy law changes that would prevent the company using personal information taken from Facebook and Instagram posts to train its AI, arguing the AI needs to learn “how individuals discuss Australian concepts”.
In a submission to the Productivity Commission’s review on harnessing data and digital technology, published this week, the parent company of Facebook, Instagram and WhatsApp argued for a “global policy alignment” in the Albanese government’s pursuit of privacy reform in the AI age.
Meta said generative AI models “require large and diverse datasets” and cannot rely on synthetic data – data generated by AI alone. The company said available databases, such as Australian legislation, were limited in what they could offer AI compared to datasets containing personal information.
“Human beings’ discussions of culture, art, and emerging trends are not borne out in such legislative texts, and the discourse that takes place on Meta products both represents vital learning on both how individuals discuss Australian concepts, realities, and figures, as well as, in particular, how users of our products engage,” Meta said.
“This means that authentic and effective learning to ultimately power meaningful products of communication is best realised from training that includes those discussions and artefacts themselves.”
Meta has been training its AI, Llama, on publicly accessible Facebook and Instagram posts since last year. The company was ordered to stop training its data on users’ posts for those based in Europe, and Meta ultimately gave users in the EU an opt-out option.
Guardian Australia reported last year that such an option was not available to Australian users because the opt-out option in Europe was “in response to a very specific legal frame”.
Meta said in the Productivity Commission submission it was “concerned that recent developments are moving Australia’s privacy regime to be out of step with international norms, impose obligations on industry that conflict with broader digital policy objectives to promote age appropriate and safe experiences online, and disincentivise industry investment in AI in Australia or in pro-consumer outcomes”.
Hardware giant Bunnings, which is appealing against a privacy commissioner finding last year against the company’s trial of facial recognition technology in select stores, also took aim at Australia’s privacy laws. The company told the Productivity Commission that while it was “committed to protecting customer privacy” it believed “every team member deserves to feel safe at work, and every customer should be able to shop without fear of harm”.
after newsletter promotion
“Privacy must be considered in the context of an employer’s strict liability and an occupier’s legal obligations to maintain a safe place of work and business,” it said.
Woolworths said it supported privacy reform but said “the proposals as presently structured could pose unnecessary challenges in how we serve our customers who increasingly expect personalised engagement and a single, frictionless shopping experience”.
Tech giant Google said there was too much regulatory uncertainty in Australia around AI, including the proposed AI guardrails, and the company repeated its call for changes to copyright law to allow its AI to be trained without breaching copyright.
