AI and Instagram Clues Led French Burglars to Canberra Luxury-Goods Target

Four French nationals who admitted breaking into a home in Canberra’s south used artificial intelligence and social-media photographs to identify their target, according to an agreed statement of facts tendered in the ACT Supreme Court.

The men pleaded guilty in March to aggravated burglary and joint commission theft after luxury watches, jewellery, rings and designer handbags were taken from the property in October 2025. The case offers a striking account of how seemingly ordinary details shared online can be pieced together to locate a person and their home.

Social-media clues led to Canberra

Court documents reported by ABC News say one of the offenders, Ilyes Abdelkader, examined Instagram images posted by one of the victims. He used artificial intelligence to assess a cherry blossom tree visible in one photograph, helping narrow the location to Canberra. Another image showed Black Mountain Tower, providing a further geographic clue.

Photographs of the home’s backyard also showed a swimming pool and tennis court. Abdelkader then used ChatGPT to search for Canberra suburbs and properties matching those features, eventually identifying the residence.

Abdelkader, Mohamed Abdelkader, Mohamed Naimi and Antoni Voisin travelled from Thailand to Sydney on October 7 before heading to Canberra. The agreed facts say they entered the home several days later while its occupants were away, using tree trunks to smash a window. They then gained access to a safe that the victims believed had been left locked.

Value of stolen goods revised

The alleged value of the haul has been contested throughout the proceedings. Police initially said the stolen property could be worth as much as $10 million. The agreed statement of facts tendered on Monday revised the provable total to just under $2 million.

Prosecutors said the authenticity of some unrecovered items could not be established beyond reasonable doubt because the victims could not provide receipts, transaction records, proof of purchase or certificates. Among the items claimed to remain missing was a rare Richard Mille RM88 “Smiley” watch, which the victims valued at $5 million.

The court documents say goods valued at about $915,000 have been returned, while items with an agreed value exceeding $1 million remain outstanding. Other stolen property had allegedly been sent to France through a freight service.

Arrests followed within days

NSW Police arrested the four men at a KFC outlet in Wentworthville, in Sydney’s west, on October 19. Investigators later recovered designer goods from a short-term rental property in the same suburb, while French authorities located other items. The four men have remained in custody since their arrests.

The ACT Supreme Court is expected to set a sentencing date next week. The guilty pleas resolve the central criminal allegations, although the dispute over the value and authenticity of particular items has affected the amount attributed to the theft.

Case highlights risks of online location clues

The case demonstrates how multiple fragments of public information can expose far more than any single post reveals. A landmark in the distance, distinctive vegetation and visible property features were allegedly combined with an AI search tool to move from general images to a specific Canberra address.

While the proceedings concern the actions of the four convicted men, the method outlined in court also underscores a broader security risk for people displaying valuable possessions online. Images can contain identifiable landmarks, reflections, street features and other details that allow strangers to infer where they were taken.

Authorities have not suggested that posting photographs makes victims responsible for offences committed against them. The alleged planning described in this case instead shows how offenders can exploit publicly available material, and why reviewing location clues before sharing high-value items or private spaces may reduce unwanted exposure.

Leave a Reply

Your email address will not be published. Required fields are marked *