Are you using AI as your go-to health expert? Here is what you need to know…
As Humans we are always looking for easier, quicker, more efficient ways of doing tasks. Time is valuable, and therefore we often look for ways to reduce the amount of time, and sometimes the effort in doing the same task but in different ways. The saying “Necessity is the mother of invention” has brought about inventions like the washing machine, the dishwasher, the computer…all things we feel are essential to our modern lifestyle. We can now devote much of our time that in the past would have been spent on manual tasks, to other pursuits.
AI (artificial intelligence) has been around for a while, but maybe not in the sense we think of it today. For example, predictive text on various devices has been around since the early 1950’s although it wasn’t until 1994 that it became used in ways we recognize more so today.
We live in a world where information is available at the tip of our fingertips. Long gone are the days of libraries, indexes and manually searching through hard copy books. For those who grew up in a time where these things were once the norm, they often give a sigh of relief. And for those who are have or who are growing up not knowing anything other than google search, social media, Apps and AI, it may be hard to understand that these things come with limitations and cautions.
We also are living in a time where people are eager to take their healthcare into their own hands. They have often seen how the healthcare system has failed them time and time again. People are tired of being ignored; gas light; dismissed; told they are being hypochondriacs, that “it is all in their head”; that the tests they are requesting aren’t necessary; misdiagnosed or not diagnosed at all; and the very thing that we all need: the human touch – empathy, understanding, and time. Instead they are given a pill to fix, cure, or treat every ailment under the sun. But nothing is ever asked of how their sleep is; what stress they are under; their diet and nutrition; their daily habits; or their exercise regime. Simply because these things take time to discuss and provide guidance on, and we all know time is money.
Enter AI. AI is designed to simulate human intelligence. Search engines are many and varied and each database has been developed by humans who spend a lot of time and money in trying to create a program that will be chosen above their competitive counterparts. Each database is therefore different, with open model and paid for versions which give you access to information not available within the public model. These databases are huge with large amounts of datasets that have been added in which the AI model can draw from to answer questions, formulate summaries, write articles and create content.
More and more people are turning to AI search engines to provide them with the answers and guidance they so desperately seek around health. They enter their genetic reports, blood test results and ask probing questions in order to get feedback on next steps and supplements to take.
The limitations and dangers of using AI as your doctor or health expert
- A lot of the information you get out of AI depends on what information was originally put in.
- The data input date for each model has a cut off date. Therefore information may be outdated, leaving out current data from the information it supplies.
- AI models are not always trained to search for answers in specific places, for example instead of searching scientific studies or review papers, AI often draws it’s answers from sources that could be incomplete or inaccurate in themselves.
- Even if you ask for references or sources of the information, it may be outdated. It can also only draw from what is publicly available. Some scientific medical studies are not available to the general public.
- AI often will generate a plausible sounding answer, that to the untrained eye, will often be taken as fact. It came from AI therefore it must be accurate. However this is not always the case. AI can make mistakes and can give our false or misleading information.
- AI models are trained according to input over time. Therefore the user often determines the kind of output they will get out.
- AI also depends on the amount of training it gets, the datasets it draws from, and on the biases it was trained in.
- It can keep incorporating the flaws if not corrected which then become part of its database and these flaws get repeated.
The things AI leaves out
When it come to genetics, AI relys on data. The genetic data you put in and the recommendations it provides will be based on that data. It will not take into account who you are as a person: your age, your background, your nutritional needs, your sleep, your stress, your past health history, your health concerns, your diet, your lifestyle, your family history, your current medications and supplements, what has worked and hasn’t worked for you in the past. It also does not always know how to interpret genotypes and phenotypes and the many gene-gene interactions that result. AI removes the human component that is often needed in providing a comprehensive feedback report based not only on your data, but on your epigenetics (all the bits and pieces that make up your unique life that shape your genetic expression).
Data remains Data without connecting the dots
Relying on AI to provide you with a supplement regime is no different to going to a doctor not trained in genetics with your genetic results. They will not know how to connect the dots between your genetic data and your current health concerns.
Here is an example. Let’s say you have 5 people with MTHFR C677T homozygous variant. They each use a different AI model and each type in that they have this gene and ask for recommendations. The information each engine gives these 5 people will be varied, but mostly it will provide the same standard answer: you need to take methylfolate because of the health risks involved.
Now let’s say each of these 5 people each have a different age group, gender, health issues, diet, lifestyle, etc not to mention the many other genes not mentioned. Yet their answer is the same. In nutrigenetics, the practitioner takes time to analyse as much genetic data of each person as possible in context with as much information they request from the person as possible to form a complex puzzle that they try to piece together. They then provide a personalised feedback report that takes into account the personality, limitations, diet preferences, mobility, age, health issues, nutritional needs, phase of life, other conditions or considerations, genotype, phenotype, stress, job, family history etc.
AI lacks what humans need the most: The human context to provide personal insights.
Why a trained nutrigenetic is still better than AI
So if you are turning to AI for health advice, for genetic recommendations, please remember that there are trained experts in this field who still take the time, care and attention to detail without relying on AI to help you improve your health who would love to help you interpret your genetic data and provide you with personalized recommendations based on you as a whole, not just your data.
Contact me for your personalized feedback report, where time, care and attention is given along with the human facet you need the most: experience and empathy, not just data.