Australian watchdog warns on deepfakes

Australia’s eSafety Commissioner says AI-powered apps that generate sexualised deepfake images are being downloaded about 200,000 times a month by Australians. Julie Inman Grant told a parliamentary inquiry that her office had taken enforcement action against eight AI-powered “nudify” services, and warned that deepfake image-based sexual abuse is increasingly being reported in Australian schools. The original report also describes concerns about voice cloning, face swapping and gaps in laws covering some non-sexual deepfakes.

For a business owner, the point isn’t only that these tools can cause harm. It’s that safeguards, oversight and responsibility for foreseeable risks are now part of the conversation about how AI products are built and used. Inman Grant said regulators need stronger powers to examine algorithms and test systems, and argued companies shouldn’t be left to assess their own risks. She also said existing rules don’t let her office test whether generative AI systems can produce some categories of illegal material. Businesses should consider where AI is used and what safeguards need attention.