Thu Sep 12 05:50:32 UTC 2024: ## Meta Grilled by Australian Senate Over AI Training Practices

**Canberra, Australia** – Meta, the tech giant behind Facebook and Instagram, faced intense scrutiny from an Australian Senate committee over its use of Australian user data to train its artificial intelligence (AI) models.

The Select Committee on Adopting AI, chaired by Labor Senator Tony Sheldon, questioned Meta representatives about the company’s practice of using public posts, including photos, from both platforms since 2007. Meta’s Global Privacy Policy Director, Melinda Claybaugh, claimed that the company only uses “public data” that users have explicitly chosen to share. However, this justification drew criticism from Senator David Shoebridge of the Greens party, who argued that Meta’s broad collection of public data effectively means all photos and text from public posts since 2007 have been scraped for training unless users made a conscious effort to set their accounts to private.

The committee also raised concerns about the potential use of data from adolescents’ accounts. Claybaugh assured the committee that Meta only uses data from users over 18, but Senator Shoebridge highlighted the ethical implications of Meta’s practices.

Despite these concerns, Claybaugh maintained that Meta takes measures to prevent personal data from being associated with individuals and mitigate privacy risks in its AI development process. She referenced the ongoing legal debate surrounding the use of data for AI training in Europe, where Meta has paused the launch of its AI products due to uncertainties over existing privacy regulations. While users in Europe have an opt-out option, Meta has not offered a similar choice to Australian users.

The Human Rights Watch (HRW) echoed these concerns, warning that Meta’s practices could pose privacy risks to children, including the potential for malicious actors to use scraped images to generate explicit content. HRW researcher Hye Jung Han urged the Australian government to introduce legislation protecting children’s data from AI misuse.

The Senate committee’s investigation highlights the ongoing debate surrounding the ethical and legal implications of using user data for AI training. Meta’s stance on transparency and user control in this context remains unclear, prompting calls for stricter regulations and greater oversight of AI development practices.

Read More