Meta urges the Australian government to refrain from implementing privacy law changes that could impede its use of personal information from Facebook and Instagram posts for training its AI. The parent company of Facebook, Instagram, and WhatsApp advocates for a “global policy alignment” as the Australian government pursues privacy reforms in the AI era, emphasizing the necessity of large and diverse datasets for generative AI models and arguing that available databases lack the depth of discussions and artifacts found in social media.
Meta argues that the training of AI without authentic data may lead to ineffective or meaningless outcomes and emphasizes the importance of understanding how individuals discuss Australian concepts, realities, and figures. The company points out that while it halted AI training on data from European users’ posts in compliance with local privacy laws, an equivalent opt-out provision is not available to Australian users due to different legal frameworks.
Meta expresses concern that Australia’s privacy laws may diverge from international norms and could negatively affect industry investment in AI and online safety measures. Hardware giant Bunnings, appealing a privacy commissioner’s ruling against its trial of facial recognition technology in some stores, also criticizes Australia’s privacy laws, insisting on the importance of safety for both employees and customers.
Google echoes the concern over regulatory uncertainty in Australia regarding AI, including proposed mandatory guardrails, and reiterates its call for changes to copyright law to facilitate AI training without infringing on copyright. The debate highlights the tension between privacy protection and technological innovation in Australia.
Source: https://www.theguardian.com/australia-news/2025/jul/17/meta-ai-facebook-instagram-personal-information-social-media-posts-learn-australian-concepts
