Deputy Transport Minister Says AI Cargo Scanners at Malaysia’s Entry Points Are Still Learning to Identify Goods

Artificial intelligence-powered scanners at Malaysia’s entry points are not yet fully effective at identifying goods because they lack sufficient data, Deputy Transport Minister Datuk Hasbi Habibollah told the Dewan Rakyat on Monday, The Star reported.

Hasbi said he had seen the problem himself on visits to sites where the scanners are installed, including Port Klang and the Malaysia Airlines cargo area. When the AI generates an image, he said, it still cannot accurately identify the items shown, which means its data is not yet sufficient. Officers on the ground are now collecting data to feed the system, in line with a recommendation from the Malaysian Anti-Corruption Commission, and the ministry is considering working with other countries to share data and strengthen the system, which he said would save time and help gather the material needed to train it.

He was answering a supplementary question from Kuala Kangsar MP Datuk Iskandar Dzulkarnain Abdul Khalid about whether AI scanners at cargo terminals can detect the smuggling of dangerous goods. The effectiveness of any such system, Hasbi said, depends on the technology, its configuration, database, algorithms and detection accuracy, as well as the skill of the inspectors using it.

What it means in practice

Hasbi stressed that AI should support inspections rather than be the sole basis for clearing cargo. Decisions must still follow established security procedures, including image assessment by inspectors, approved inspection methods, further checks where needed and compliance with approved security programmes. AI can help inspectors spot anomalies, he said, but any follow-up action must remain subject to verification by qualified personnel. The admission is a useful reality check as the government expands AI at the borders, including the AI-aided national border control system the Home Ministry says is underway, and as Parliament prepares for the AI governance bill expected in the Dewan Rakyat early next year.

Scanners that learn from cargo images improve only as fast as labelled data arrives. What to watch is how quickly accuracy improves, which countries Malaysia approaches for data sharing, and whether the government publishes performance figures for the AI systems at its ports and airports.