Artificial intelligence has been blamed for many things: displacing copywriters, writing mediocre code, and occasionally hallucinating bizarre historical facts. Until recently, directing a foreign burglar to a high-end stash of luxury goods in suburban Canberra was not on the bingo card.
Australian courts are currently processing a bizarre case involving a French national who used ChatGPT as a digital scout. By feeding prompts into the large language model, the perpetrator was able to analyze geographic and demographic data in the nation's capital to identify properties likely to house valuable items. It reads like science fiction, but it is a sobering milestone in the intersection of generative AI and conventional crime.
How the Digital Heist Unfolded
The details emerging from the investigation highlight the surprising versatility—and vulnerability—of publicly available AI models. Large language models are trained on vast oceans of public data, including real estate listings, council development approvals, social media footprints, and local news archives.
While safety guardrails built into tools like ChatGPT generally prevent users from asking 'how do I commit a crime,' clever prompting can often circumvent direct restrictions. By asking abstract, analytical questions about neighborhood demographics, high-value suburbs, and architectural styles associated with wealth, the user effectively reverse-engineered a target list.
It requires no specialized hacking skills or dark web credentials. It simply requires a user who understands how to query a public intelligence engine to aggregate data that was already sitting in plain sight on the internet.
The Open-Source Intelligence Problem
This Canberra incident shines a harsh light on Open-Source Intelligence (OSINT)—the practice of collecting information from public sources to use in analysis. Historically, OSINT was the domain of investigative journalists, financial auditors, and intelligence agencies.
Today, generative AI has democratized OSINT. Anyone with a smartphone can aggregate millions of data points in seconds. Property listings show floor plans and expensive interior finishes. Local community groups on social media discuss who is away on holiday. Council DA applications reveal major renovations and high-end security upgrades.
Every public photo of a renovated kitchen, every check-in at a holiday destination, and every trade invoice shared online contributes to the digital profile that AI tools can scrape, aggregate, and analyze.
Law Enforcement and AI Safeguards
The case has reignited debates among lawmakers, cybersecurity experts, and tech developers regarding the ethical boundaries of AI. Tech companies constantly update safety filters, but the underlying technology is designed to be helpful, informative, and broad-reaching.
When a tool is powerful enough to summarize legal documents or write computer code, it is often powerful enough to correlate disparate pieces of public information into actionable intelligence. Law enforcement agencies globally are scrambling to develop frameworks to prosecute not just the physical perpetrators, but the digital tools used to facilitate crimes.
What This Means for the Physical Security Paradigm
For years, home security focused on physical barriers: deadbolts, sensor lights, and security cameras. As crimes become increasingly data-driven, physical security must be paired with digital hygiene.
High-net-worth individuals are not the only ones vulnerable to AI-assisted targeting. Small business owners, contractors, and tradies often store valuable assets—tools, materials, and equipment—in vehicles, sheds, and commercial yards. If public permit filings, business registries, and social media posts map out where expensive gear is housed, bad actors can optimize their routes just as easily.
The Tradie Angle: Securing Your Assets and Client Data
How does a Canberra burglary involving ChatGPT affect an electrician in Brisbane or a plumber in Perth? The connection lies in how trade businesses handle digital footprints and physical security.
Tradies are frequent targets for theft because work vans and sheds are packed with expensive, portable assets—laser levels, cordless kits, copper piping, and specialized diagnostic tools. When tradies broadcast their locations, client addresses, and project timelines online, they inadvertently create a roadmap for opportunistic criminals using modern search tools.
- Limit public asset displays. Avoid posting photos of newly delivered high-end commercial tools or expensive material stockpiles with geotags turned on.
- Protect client privacy. Never share specific client addresses or high-value residential project locations on public social media channels.
- Secure your business records. Ensure customer databases, invoicing histories, and job site notes are kept in secure, encrypted platforms rather than floating in unsecured email threads or paper pads left in vans.
Keeping Your Business Operations Locked Down
Running a trade business requires trust. You enter clients' homes, handle their keys, and collect sensitive financial details. Protecting that data is just as important as locking your toolboxes at the end of the day.
This is where modern software makes a tangible difference. Instead of keeping client lists, asset inventories, and financial data scattered across random notebooks, unsecured text messages, or vulnerable desktop folders, trade businesses need centralized, secure infrastructure.
Dockett was built specifically for Australian sole traders and small teams who want to run a tight ship. By streamlining voice-to-invoice, benchmarked pricing, and client re-engagement in a secure, cloud-backed app, Dockett keeps your business data organized and protected—leaving you to focus on the tools of your trade, without worrying about digital blind spots.