NSW criminalizes political deepfakes ahead of 2027 elections

Deepfakes have been deployed to influence elections in countries such as India, Mexico, Ukraine, the U.S. and Taiwan, prompting some countries to introduce laws that combat political deepfakes. New South Wales (NSW) recently became the second Australian jurisdiction after South Australia to legislate against political deepfakes.
The government framed the reforms as a safeguard against AI-driven misinformation ahead of the 2027 state election, according to a new research paper from the Parliament of New South Wales. The report details potential adverse impacts of political deepfakes, summarizes legislative reforms, and discusses further measures.
Existing NSW and Commonwealth laws already cover sexually explicit deepfakes, but until now, no state law has addressed the broader category of political deepfakes. In March, the state passed the Electoral Legislation Amendment (Elections) Act 2026, introducing two new offenses under the Electoral Act 2017.
The first prohibits publishing or distributing digitally generated material that depicts a real person doing something they did not do, while the second prohibits publishing realistic AI-generated depictions of people, events, or places unless the material is clearly labeled as digitally created. Both offenses apply only during an election period.
Penalties for the first offense reach up to AU$6,600 (US$4,790) and up to six months imprisonment for individuals, and AU$33,000 ($23,960) for organizations. The second offense carries lower penalties of up to AU$2,200 (US$1,600) and up to six months imprisonment for individuals, and AU$11,000 (US$7,990) for organizations.
The legislation drew constitutional scrutiny during parliamentary debate. The Opposition argued the new offenses could infringe the implied freedom of political communication and suggested amendments. The government rejected the amendments, asserting the laws were sufficiently targeted to avoid chilling legitimate political speech, satire, or parody.
The research paper also examines other countermeasures against political deepfakes under consideration by the NSW government. These include watermarking, automated detection tools, platform-level takedowns, and public education. The NSW Electoral Commission ran a digital literacy campaign called ‘Stop and Consider’ for the 2023 state election.
The government is also attempting to work with tech platforms. During Australia’s federal elections in May 2025, Meta pledged to remove deepfakes and other forms of disinformation.
How the new NSW regulation performs in practice will become clearer when the state goes to the polls in 2027, the report concludes.
“The 2027 NSW election will provide an opportunity to see whether there are any cases of political deepfakes, what their impact will be, how the new laws are applied and whether additional measures are required to safeguard elections in NSW,” it notes.
Article Topics
AI fraud | Australia | deepfake detection | deepfakes | elections | New South Wales (NSW)







Comments