The Australian Senate today (24th) reviewed a new bill introduced by independent senator David Pocock, which would prohibit the non-consensual sharing of AI-generated deepfake images, with offenders facing heavy penalties and potential civil lawsuits. Pocock said that aside from sexual deepfake material, Australia currently has “virtually no protections” against general deepfake content, and that the government has been “slow to react” to rapid AI developments.
Under the bill, individuals who share non-consensual deepfake material could face fines of up to AUD 165,000. Companies that fail to comply with removal orders from the eSafety Commissioner could be fined up to AUD 825,000. Offenders who ignore takedown notices may also face a second penalty. The bill also proposes amending the Privacy Act to allow victims to directly file civil claims against offenders for damages without needing to prove specific financial loss, reflecting the emotional harm caused by deepfakes.
Pocock stressed that a person’s face, voice, and mannerisms are important markers of identity and should be legally protected. The bill includes exemptions for news reporting, satire, law-enforcement purposes, and situations involving children, and aims to avoid penalizing those who share content unintentionally, while still targeting those who act knowingly or recklessly.
Pocock and independent MP Kate Chaney criticized the government for being “absent” on AI regulation, noting that current laws only prohibit sharing sexual deepfakes and fail to address other forms of AI misuse. They urged the government to urgently develop a comprehensive AI strategy and establish an Australian-based AI safety research body.
The government stated that it will not rush to introduce AI legislation before the completion of a related review expected at the end of 2025. But Pocock warned that AI risks are like a “runaway train,” and that without timely regulation, Australia may be unable to cope with the impacts of technological change.