Study Reveals AI Can Spot Toxic Food Before It Reaches Consumers

(MENAFN) Artificial intelligence may soon prevent millions of deaths caused by food contamination, thanks to a breakthrough led by Australian researchers.

A new study, released Tuesday by the University of South Australia (UniSA), reveals that pairing hyperspectral imaging (HSI) with machine learning offers a rapid and precise way to detect mycotoxins—dangerous compounds from fungal contamination—in grains and nuts during critical stages such as growth, harvest, and storage.

These toxins are strongly linked to cancer, immune system suppression, and hormonal imbalances. According to the World Health Organization, foodborne contaminants like mycotoxins are responsible for 600 million illnesses and 4.2 million deaths globally each year.

Existing detection techniques are often too slow, expensive, and destructive for real-time use in large-scale food processing, explained UniSA PhD researcher Ahasan Kabir, the study’s lead author. The findings were published in the journal Toxins, based in Switzerland.

"In contrast, hyperspectral imaging -- a technique that captures images with detailed spectral information -- allows us to quickly detect and quantify contamination across entire food samples without destroying them," Kabir said.

The HSI system works by analyzing spectral "footprints," and when integrated with machine learning, it can pick up subtle spectral shifts that signal contamination. Kabir conducted the research with colleagues in Canada and India.

Their review of over 80 recent studies demonstrated that machine learning-enhanced HSI consistently outperformed traditional testing methods, especially in identifying aflatoxin B1—among the most potent known foodborne carcinogens.

According to the researchers, the technology is versatile enough to be mounted on industrial processing lines or adapted into handheld devices to scan crops such as wheat, maize, peanuts, and almonds. They are now fine-tuning the approach to boost precision and consistency.

MENAFN12082025000045017169ID1109917049