ON the heels of widespread concern about the presence of mind-altering drug components in some children’s snacks, Adrian Dunkley, founder of StarApple Analytics Limited — a local artificial intelligence (AI) company — says technology can play a role in eliminating that threat.

In a statement to the media on Thursday, following reports that another children’s snack has surfaced containing THC (Tetrahydrocannabinol), the substance primarily responsible for the effects of marijuana on a person’s mental state, Dunkley in describing the situation as “grave” said there is need for a multidisciplinary approach with AI playing a major role.

According to Dunkley, multi-modal AI models working alongside public health officials and everyday Jamaicans is part of the solution.

“A multi-modal model is designed to understand, interpret and generate insights based on multiples types of data including, but not limited to text and images,” Dunkley explained.

He said the capacity for varied data analysis makes the model “uniquely suited for complex tasks such as such as differentiating harmless candies from harmful counterparts using just a photo”.

According to Dunkley, StarApple AI, in a recent research to address the issue of foods harmful to children disguised as ordinary candy, tested 24 popular candies brands, across various scenarios, angles and backgrounds, aiming to discern which ones contained THC or could be harmful to children.

Dunkley said the research conducted by Section 9, a specialised research group in StarApple AI focusing on Social Good and Ethical AI, utilised GPT-4V, a multimodal model.

The StarApple AI founder said the model, from the input question “is this candy safe for kids” in respect of the THC candy which was reported to have hospitalised more than 60 Jamaican students recently, responded with “no this candy is not safe for children; the packaging indicates that it contains THC which is the psychoactive compound found in cannabis. The packaging also states that it contains 1,000 mg of THC which is a very high dose even for adults; always keep such products out of the reach of children”.

The response was virtually the same for two other “lookalike” candy brands which were researched, Dunkley said.

The StarApple AI founder is, in the meantime, emphasising that utilisation of the multimodal model isn’t confined to identifying unsafe candies.

“It is inherently adaptable and can be further trained to analyse packaging designs for compliance with local regulations, generate safety recommendations, and more. As AI technologies continue to mature, their applicability in safeguarding public health can only be expected to grow,” he stated.

“We are at a critical juncture where technology and public health intersect. This study signifies a step in utilising AI for real-world applications that can have immediate and long-lasting impact. With continuous advancements in scientific techniques, AI’s role in enhancing public safety is not just promising—it is inevitable,” Dunkley added.

He noted that incidents of THC poisoning due to misleading candies have been reported since 2019 in multiple countries, including the United States, Canada, and the United Kingdom.

Said Dunkley: “The integration of AI models like the one used by Section 9 can serve as an invaluable tool in the toolbox of public health workers, citizens, and regulatory bodies, mitigating the risks associated with purchasing and consuming unsafe products.”

On Monday, October 2, more than 60 Ocho Rios Primary School students were rushed to St Ann’s Bay Hospital after consuming sweets laced with THC. At the end of that day at least seven were still hospitalised. Since then a vendor identified as Devon Holness of Mansfield Heights, Ocho Rios in St Ann, was named as a person of interest in relation to the sale of ganja-laced sweets.

Also, following the incident, the National Council on Drug Abuse (NCDA) said the affected students will be closely monitored over the next few weeks for psychological effects. It further said it would be conducting tests on the sweets to determine what specific drugs were infused.