Papua New Guinea criminalizes AI deepfakes in Cybercrime Code reform
Government moves to criminalize AI-enabled fraud
Papua New Guinea will amend its Cybercrime Code Act to criminalize fraudulent voice cloning, deepfakes, and digital impersonation, the government announced in August 2026. The Department of Information and Communications Technology, working with the National Information and Communications Technology Authority and Department of Justice, is drafting amendments targeting what officials describe as "emerging harms" from the non-consensual use of AI and digitally manipulated identities ABC News.
Acting Information and Communications Technology Minister Peter Tsiamalili Jr. characterized the reform as a priority response to AI-driven fraud. The government opened public consultation on the draft amendments beginning 6 August 2026 ICT PNG, with Parliament introduction targeted for the November 2026 sitting.
Scope of criminalization
The proposed amendments explicitly target fraudulent voice cloning and digital impersonation as prohibited conduct under the reformed Cybercrime Code. The government has not yet disclosed specific penalty terms, fines, or prison sentences in publicly available announcements, though the consultation process invites stakeholder input on enforcement frameworks CoinGeek.
The timing reflects broader regional concern over AI-enabled identity abuse. Similar consultations on AI governance are underway across the Pacific, with governments weighing regulatory approaches to deepfake production and unauthorized voice synthesis.
Why this matters for the agent economy
The PNG reform targets infrastructure that bad actors—including autonomous agents and manual operators—use to execute scams and fraud at scale. Voice-cloning technology and deepfake generation have become commoditized through both commercial APIs and open-source models, enabling low-friction impersonation attacks. Criminalization of the *use* of these tools for fraud (rather than the tools themselves) reflects a jurisdictional strategy: make the conduct illegal, leaving legitimate applications intact.
For agent developers and deployers in PNG and neighboring jurisdictions, the legislation signals that identity spoofing—whether by autonomous systems or humans—will carry criminal liability. This may accelerate adoption of identity-verification layers and consent-based voice/image use in agent design.
The reform also creates a legislative precedent in the Pacific region, potentially influencing policy in Australia, New Zealand, and other neighboring nations wrestling with similar harms.
Next steps
Public consultation runs through the coming weeks, with a target introduction date of November 2026. The government has not announced a public hearing schedule or detailed penalty structure as of mid-August 2026.