Much of the debate has taken on cataclysmic overtones, with Elon Musk recently branding AI “one of the biggest threats to humanity” while Geoffrey Hinton, one of the so-called “godfathers of AI” has warned that the technology could “take over” if we lower our guard.
In early November, the UK hosted an AI Safety Summit which resulted in a landmark international agreement between world powers including the EU, US and China to rein in this rapidly-developing technology. Targeting “frontier AI models,” such as OpenAI’s ChatGPT and similar generative AI programmes, this regulatory attempt arrived just days after the Biden administration and G7’s own moves to mitigate AI risks.
Given its immense potential for misuse, AI unquestionably requires greater regulation, yet the understandable focus on its dangers risks drowning out its enormous potential to improve our lives. Government policies should therefore be mindful to promote innovative applications of AI in beneficial areas, such as public health, where AI technology is fueling the development of personalised nutrition as an accessible, future-fit approach to dietary health.
Between outdated nutritional guidance and one-size-fits-all fad diets on social media, consumers need access to reliable dietary information. In recent years, innovative AI-driven companies such as basys.ai, Youniq Health and Nuvilab AI have emerged with dietary solutions tailored to individual health needs and aimed at improving global health outcomes.
Basys, for instance, tackles chronic conditions like diabetes by working with medical insurers to easily process burdensome authorisation requests, while Youniq and Nuvilab are emblematic of a new range of AI-powered nutrition apps. Instead of recommending a traditional diet, including which foods to avoid, these apps function more as personal assistants to help users make healthy eating choices. The apps are based on research that bodies react differently to similar foods, and that the healthiest food is likely to be unique to each person.
Youniq can process a range of personal information, from blood samples, genome tests, even blood pressure and height and run it through an algorithm that will suggest recipes tailored to a specific individual. Helpfully, the algorithm is also able to scan the ingredients that users have in their fridge and use the data to compile simple recipes. Nuvilab, in turn, is focused on people who need to accurately record their diets, such as patients with chronic diseases and senior citizens in care facilities. Its technology notably improves the accuracy of dietary recording and can analyse dietary habits to provide reports on the nutritional intake rate of each meal.
In the United States, the AI nutrition boom has begun to spark serious research interest. The National Institutes for Health recently announced a research programme focused on developing algorithms that predict individual responses to food and dietary patterns. The programme argues that there is no such thing as a perfect, one-size-fits-all diet, and that advances in AI can help provide insights in personalised nutrition.
The picture across the Atlantic is less positive where, instead of focusing on the future of nutrition, the European Union is still caught in a longstanding argument over nutrition labels. While the US pours money into AI solutions, the EU is still considering options for its nutrition label proposal, despite the France-backed Nutri-Score label at the centre of the debate attracting heavy criticism due to its overly-broad, outdated and scientifically-flawed approach.
Summing up the fundamental problem, renowned French nutritional expert Jean-Michel Lecerf has commented that “Nutri-Score is…a nutritional concept from the 1980s” and “offers a reductionist approach” to improving dietary health and tackling obesity in 2023 –particularly in light of updated science on fats and metabolic syndrome. Citing the health risks that Nutri-Score’s misleading algorithm poses to consumers, Romania and Italy have gone as far as to ban the label.
Meanwhile, opposition is mounting across Europe, with Switzerland’s parliament mulling over a potential Nutri-Score prohibition over concerns that the label does not provide shoppers with sufficiently-detailed information – which creates a worrying gap for people with specific health and dietary needs. With AI-powered apps likely to only grow in popularity as users look to technology for answers to the nutritional puzzle, Brussel cannot let a dead-end labelling debate leave Europe in the dust.
Considering that many AI-based nutrition apps are seemingly not yet ready for widespread adoption, further EU procrastination on this front could have real consequences. Indeed, the industry is still in a “Wild West” phase, with nutritionists warning users to be cautious of scientifically-dubious or even fraudulent companies.
Moving forward, AI-based nutrition developers and medical professionals will need to work closely together to implement ethical AI policies, such as those recommended by the World Health Organisation (WHO), to mitigate potential privacy concerns and unintended consequences. The WHO has warned against a rush to implement untested, or poorly-tested, AI systems, given that technologies which are adopted too soon can cause errors by health-care workers, harm patients and erode trust in AI, thus undermining their potential benefits for public health.
While certain fears are well-justified, AI is clearly not going anywhere. Indeed, we are only just beginning to grasp its transformative potential. With the right balance between regulations to protect citizens and spur innovation, personalised nutrition can have a transformative global impact on public health. Off the back of the UK’s historic AI summit, as well as the bold policies emerging from the US and G7 in recent weeks, political leaders in the EU and around the world should take inspiration and focus on the future rather than staying trapped in the past.
