Public choices / October 2026

It sounded like someone you trust.

A familiar voice, polished document or convincing video can be manufactured. The response needs safer platforms and payments as well as a chance to stop and check.

Open the room All Twinkles Twinkle 35 · video

Twinkle / policy room

Connected Twinkles

Twinkle 35 / Brainstorming scenario

AI scams: responsibility across the whole chain

Make impersonation fraud a platform, telecom and payment-system responsibility as well as a literacy issue. Provide a supported route from report to investigation and recovery assessment.

A convincing voice or video is not proof of identity. The connected idea is practical verification plus responsibility across the services that spread the scam and move the money.

Explore the connected work

AI Trust Index Australia - Practical AI literacy and impersonation-scam explanations, alongside official reporting and support routes. Source repository
Relevance: Australian law - The source-led method for following legal threads and understanding how laws touch an ordinary life. Source repository
Australian Legal Engine - Offline code that parses, indexes and retrieves quoted statutory provisions with source citations. It does not decide a legal case. Source repository

How it could work

Participating services verify advertisers and maintain secure escalation for impersonation, with rapid review of repeat offenders.
Payment providers offer accessible warning and trusted-contact choices, with human review when an unusual transfer is paused.
A shared case reference connects reports and dispute responses so victims do not start from scratch at every organisation.

Everyday example

Someone hears a convincing cloned voice; an independent callback and a supported bank check interrupt the transfer before it leaves.

Who could act and how it could be funded

National regulators and relevant providers need coordinated standards; states can improve victim support and public procurement. Service operators fund prevention and response; any public navigation programme has a separate budget.

Fair-go safeguards

No automatic reimbursement promise or blame directed at victims. Trusted contacts are optional; safeguards must not give a coercive person control of another adult's money.

What would show progress

Fraud interrupted; time to human response; complaint handovers; justified recovery decisions; false-positive harm.

Twinkle 35 · video Open the full room Rabbit-hole policies

Full room: ideas, tools and sources

Delivery roles: Australia's three levels of government · Victoria's parliamentary responsibilities.

What we can check

Scamwatch describes AI-enabled impersonation, voice cloning, false endorsements, fake sites and personalised messages. Visual polish or a matching voice does not establish identity. Verification through an independently found official number or a known contact is more useful than trying to spot every visual flaw.

Scamwatch: AI and impersonation

Choices and tradeoffs

Platform and advertiser accountabilityVerified advertisers, faster takedowns and repeat-offender controls can interrupt distribution. Appeals and transparent error reporting are needed when legitimate speech or businesses are caught.

Payment checks and accessible disputesBetter warnings, trusted-contact options and recovery processes can reduce losses. Delays and false positives have costs; no page should promise automatic reimbursement.

Practical verification and supportTeach independent callback and trusted-source checks without blaming victims. Public education cannot carry the entire burden of sophisticated organised fraud.

A practical public voice

If money or account details are at risk, contact your bank through its official route promptly. Use Scamwatch to report and follow its recovery guidance. Keep evidence privately and beware of a second scam promising to recover funds for a fee. For policy, ask who must prevent, respond and resolve disputes across the platform-telecom-payment chain.

Recognising an impersonation

The AI Trust Index Australia guide describes AI-voice and fake-video scams. Use it as a literacy resource alongside the official reporting and recovery routes above; it is not an authentication service or scam detector.

AI scam literacy guide

Status and sources

Checked 3 October 2026. Source dates and jurisdictions remain attached to the evidence. These are choices for public discussion, not an adopted party platform.

Twinkle 35: AI scams

A question for the workbench

“Don’t trust the voice. Check the person through a different door.”