# Google's AI Training Shift Impacts Australian Data Privacy Default settings now use work-related media for AI training unless manually opted out. - **Source:** [Australia Archives - TechRepublic](https://www.techrepublic.com/topic/australia/) - **Published:** 2026-07-09 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/5799bcda-1d08-4737-83d7-9a3a38d4f90e ## Ubiquitous Google tools Google Search, Lens, and Translate are deeply embedded in Australian workplaces, used routinely for tasks like error troubleshooting and document translation. Unlike optional AI tools, these services are default work solutions on both company and personal devices. This widespread integration means policy changes affect nearly all employees automatically, regardless of individual adoption choices. ## Expanded training data scope Google now uses screenshots, photos, voice queries, and documents from daily work interactions to train AI models by default. This represents a shift from training on public web pages to utilizing employee-generated content. Users must manually disable this setting, which remains active across 88% of Australia's search market and over 90% of mobile searches according to Statcounter. ## Regulatory timing gap Australia's Privacy Act will require disclosure of automated decision-making data uses starting December 2026, with formal guidance due September 2026. Meanwhile, Google's change precedes these regulations, creating an interim period where default settings operate without specific AI governance. The government has opted against standalone AI laws, relying instead on existing privacy and consumer protection frameworks. ## Enterprise governance challenge The setting creates corporate data risks as work-related content—including customer information, contracts, or source code—may be captured through routine employee actions. Disabling future collection doesn't remove already saved data. This transforms individual privacy choices into organizational governance issues, particularly for sensitive business materials uploaded through consumer-facing tools. ## Industry-wide pattern Google follows Meta's 2024 precedent of using Australian user data for AI training without local opt-out options. The approach mirrors broader tech industry practices: default data collection, opt-out rather than opt-in mechanisms, and reliance on users not reading privacy policies. This creates consistent challenges for Australian businesses managing employee tool usage. ## Dual-use reality 13.6 million Australians used AI tools like Gemini and Copilot in Q1 2026 per Roy Morgan, with enterprise and consumer versions often operating simultaneously. Google's change introduces another consumer surface that captures work content across managed and unmanaged devices—precisely where enterprise AI governance typically fails to reach. ## Practical risks The danger lies in habitual use: employees routinely solve work problems through Google tools, with each action potentially retaining media under new settings. IT teams cannot assume policy compliance equals actual opt-outs, as the default remains active until manually disabled. This creates invisible data exposure points throughout organizations. ## Proactive measures needed Organizations must act before September 2026 guidance, implementing data classification and staff awareness programs. The Privacy Act's technology-neutral approach means defaults will continue testing legal boundaries. The burden falls on employers to manage risks from tools deeply embedded in workflows but outside traditional IT governance structures. ## Sources covering this story - https://www.techrepublic.com/article/google-ai-training-australia-apac/ — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._