AI Election Deepfakes in Australia
Parliamentary evidence on AI and elections highlights limits on what regulators can enforce — and why publishers and sharers need their own verification steps.
Short answer
Australian electoral law does not guarantee that every misleading political video can be removed or corrected by the Australian Electoral Commission — enforcement powers are specific, and many fabricated clips may fall outside them. Before sharing political video, check the original source, authorisation labels, and official AEC guidance on enrolment and voting.
A striking political video appears on a phone. It seems to show a public figure saying something outrageous. Within minutes, people are arguing about it and sharing it. But who is responsible for checking whether the clip is real?
The question became more pressing after evidence given to a parliamentary committee on 14 October 2026. Australian Electoral Commissioner Jeff Pope warned that AI could amplify misinformation and other existing threats to elections. An Australian Associated Press report published the next day also described limits on what the Australian Electoral Commission can investigate or decide.
During the hearing, members raised hypothetical examples of fabricated political videos. These were illustrations of possible problems, not reports that the particular events had occurred. The commissioner explained that the law is not a general guarantee that every political claim must be true. The commission can act in specified circumstances, including certain misleading information about how a voter casts a ballot, and it can enforce relevant rules about identifying who authorised election communications.
That leaves an important gap between content that seems misleading and content the election regulator has legal power to remove or correct. A video may be false or manipulated without falling within the commission's particular enforcement powers. Conversely, satire and legitimate political criticism should not automatically be treated as unlawful merely because technology was involved.
The practical risk extends beyond political parties. Community organisations, publishers, employers, and ordinary users may share a dramatic clip before discovering that it was altered. A correction later may reach fewer people than the original post.
What to do
Before forwarding a political clip, find its original source. Check whether the full speech or interview exists and whether established reporting confirms the incident. Look for the authorisation information required for relevant electoral communications. A missing label is a reason to investigate, not proof that a video is fabricated.
For information about enrolment, polling places, ballots, and the voting process, go directly to the Australian Electoral Commission's AI and elections guidance. Treat messages claiming a new voting rule with particular caution. If a clip cannot be verified, do not present it as established fact.
Organisations that publish political or public-interest material should introduce a short review process: identify the source, check the context, record who approved publication, and label artificial or altered material clearly where appropriate. Review obligations should account for electoral law and the importance of lawful political expression.
Sources
- Michael West Media / AAP, 15 October 2026 — parliamentary committee reporting on AI and elections.
- Australian Electoral Commission — AI and elections.
- AEC — Authorising electoral communications, updated 13 October 2026.
- ISO/IEC 42001 explained — International Organization for Standardization.
We help publishers and organisations build practical checks for AI-altered content, approval records, and correction processes — aligned with ISO/IEC 42001 risk management through our gap assessment.