TL;DR: As Scam Prevention Frameworks roll out, telecom and financial institutions face a massive operational challenge: complying with extensive obligations across six principles while adversaries use automated tooling to operate at machine speed. In this post, I examine where AI, ML, and autonomous agents genuinely fit across the framework, and why their success depends on wrapping probabilistic systems inside strict deterministic boundaries.

We have been spending a lot of time whiteboarding compliance architectures for the new Scam Prevention Frameworks (SPF).

TL;DR: When my family and friends picture scam prevention, they assume a single magic box labelled “Filter”. But as technologists working under Australia’s new Scam Prevention Framework (SPF), our job is breaking that assumption down into 6 messy, connected boxes—from the customer journey at a retail counter to cross-sector intelligence and victim restitution.

I haven’t written a post in a few weeks.

The truth is, the last month has been completely all-consuming. I’ve been deep in the trenches, whiteboarding and dissecting the latest SPF Codes and Rules published by Treasury.

This is a true story.

A friend’s wife received a call from someone claiming to be from the ATO. The caller was aggressive and convincing. He told her that her husband had been arrested for tax fraud and, because of the “seriousness” of the case, the family was now under government surveillance.

Then came the demand: pay $5,000 to secure his release.

She was terrified and only moments away from sending the money.

Scams are real, and they’re everywhere#

Who hasn’t heard of scams? Many of us have. For most of the lucky ones, it’s just news we’ve heard. It’s not close or personal. But they are very real - In Australia in 2024, people lost $2.7 billion of their hard-earned money to scams.

What exactly is a scam?#

Legally, scams and fraud are different terms. While fraud is broader, scams typically involve two scenarios: