Garbage in, garbage out: Artificial intelligence could become ‘crucial’ nutraceutical industry tool if 'care is taken'
The use of AI in nutraceutical innovation has still been relatively limited till date, but this is expected to change rapidly over the next few years.
“AI is creating a new base level of operational excellence and new benchmarks for operational costs – so existing businesses that do not adapt and integrate AI quickly into their operations risk getting left behind,” industry consultant Ryan Sproull, speaking for nutraceutical consultancy firm 6AM Agency, told the floor at the recent Vitafoods Asia 2024 event in Bangkok, Thailand.
“Although its direct use in innovation is still limited, this is largely because data is still being gathered – for example, we see AI platforms such as Nutrify Today gathering information from research literature, plant genomes and traditional medicine to apply the identification of hidden patterns, bioactive properties and therapeutic applications.
“So from a practical perspective, this AI is finding compounds that have a higher chance of meeting safety and efficacy criteria, which in turn speeds up the discovery and approval process for new product development.
“This sets a new benchmark for how quickly companies can turnaround from research to commercialisation, and there is no doubt that the first adopters have the advantage but soon, once it is established as a benchmark, it will be all relevant firms that are adapting to use such tech.”
Another example of an AI platform being developed for the nutraceuticals sector is Apex Compliance, which specifically looks at the regulatory aspects of nutraceutical product development to maximise supplements’ chances of approval.
“Apex Compliance works by scanning the nutraceutical products’ websites and marketing communications for potential areas of risk and red flags that regulatory agencies would pick up on,” he said.
“Once these have been identified and removed, the platform then uses generative AI to suggest lower-risk alternative phrasing, and in fact the algorithm will learn over time how to generate better messaging as more users engage with it.”
That said, Sproull also offered a warning regarding the use of AI, highlighting that this is tech that can learn both the positive and the negative depending on what data it is fed.
“AI learns what you tell it to learn, so you must teach it the right thing with the right data – it is very much a case of garbage in, garbage out,” he said.
“So for example if there is a consistent human bias in the case studies that it is given, it will learn those biases and this will mean poor outcomes and results – for instance, if there is a consistent racial or gender bias, even an implicit one, in clinical studies data provided, this will be picked up and learned by the AI which could then make erroneous conclusions e.g. that women cannot use a promising new ingredient.”
Factors to consider
Apart from innovation and regulation, AI is also expected to make a significant different in the daily operations of food and nutra businesses simply by removing the menial and repetitive tasks.
“Business improvement with AI can come in the form of increased value, decreased costs and improved employee experiences, in situations where there is a lot of data available, many decisions are being made and feedback is rapid,” he said.
“Some examples include quality control using cameras instead of the human eye, cutting down the time needed for preclinical R&D tasks, and improving digital marketing.
“That said, for the latter we also need to remember that AI learns from previous examples and creates based on these as well, so the likelihood at this stage is for it to make more of the same as opposed to thinking creatively – which basically means that effective marketing which still calls for newness and differentiation is still very much dependent on human creativity.”
