ChatGPT, the artificial intelligence tool was trained using a diverse and extensive corpus of text from the internet pre-2021. From there it was refined into a static tool for interacting with its knowledge. If you have tried ChatGPT you may have experienced its ability to just make things up, even drawing conclusions that don’t quite make sense. But, it’s young and doesn’t have access to live information, so let’s call this a first step into the possibilities for the way we interact with shared knowledge online.
Another very interesting AI model was developed by a team at Massachusetts Institute of Technology (MIT) and is an image based tool for detecting lung cancer in x-ray scans. The model, called Sybil was trained on hundreds of labeled CT scans with visible cancerous tumors to ensure it would be able to accurately assess cancer risk. By using this detailed method, the model was able to develop some predictive capabilities as to which lung was most likely to develop cancer.
Data is essential to developing algorithms. Even more labeled x-ray images would be required for cargo due to:
Complexity of cargo shipments including mixed commodities
Identification of thousands of goods and how they look on an image
Inclusion of shipment data from text resources
With an integrated inspection program collecting data from more trade, it is possible to collect this information and have a reliable library to build detection algorithms for Customs, accelerating secure trade.
Take a look at the S2U website’s special feature, Global Smuggling Trends Report . This interacti...
Algorithms that build upon a weak data foundation can lead to subtle vulnerabilities in the screening process.
CertScan IQ performs a quality check on data to maintain the integrity of fast moving inspection operations.