Google’s voluntary 2020 action to restrict food and beverage advertising through online channels in the EU and UK would limit online advertising of the most unhealthful products to children in the United States if the policy were applied to the U.S. Market. Researchers from New South Wales University, the University of Wollongong, and the University of North Carolina came to this conclusion after using nutrient data from NIQ Label Insight, for the top 25 US food and beverage manufacturers. The researchers examined each product against four nutrient profile models (NPMs): Google’s, the World Health Organization’s Europe model, the Pan American Health Organization model, and the Chilean Government NPM. Under Google’s NPM, 18% of the 14,188 products were eligible to be advertised to children, a figure representing $44 billion out of more than $240 billion USD generated in annual revenue for the top 25 U.S. manufacturers. Researchers concluded that the effectiveness of Google’s policy would be strengthened by refining the Google NPM to better align with those developed by health authorities, including WHO.

Report: Could Google help curb online advertising of unhealthy foods to US children?

Consumption of ultra-processed foods has gone up in most countries, and high consumption of such foods (UPFs) has been associated with non-communicable diseases. Researchers from the University of Toronto and Fundación Interamericana del Corazón Argentina used NIQ data to discover that automation can help make this process simpler and faster. The researchers used food product information from the USDA Branded Food Products Database, which includes NIQ Label Insight data, as well as data from FLIP Canada, FLIP Latin America and the Caribbean to train and test a language model that could predict food-processing levels. Text information from food labels was used to fine-tune the model. The fine-tuned model achieved the highest F1 score, 0.98, in predicting food-processing levels. It outperformed other models, such as structured nutrient facts and bag-of-words. The researchers concluded that this automation strategy is effective and can speed up monitoring and regulation of ultra-processed foods in the global food supply.

Report: Accelerating Classification of Food Processing Levels Using a Fine-Tuned Language Model: A Multi-Country Study