Let’s start at the end
Fast forward just 10 years from now and what kind of world could we be living in? Based on current population growth the planet will be tipping over 8.5 billion people as we have just broken 8 billion in 2022. This number seems extraordinary when you consider there was just 4.6 billion in 1982 (do you remember the simplicity and excitement of the 80’s?)
Consider the growing city of Melbourne, Australia that is fast on track to become more populated than Sydney within the next decade and by doing so will break 6,000,000 people increasing its density from approximately 510 people per km2 to 540.
This will mean an increase of people living, moving and using resources within this city alone of 1,000,000 so an infrastructure overload is a potential barrier for the topic of this post, healthy living.
Known unknown error
There is a common phrase that “fear is of the unknown”. I use it a lot as the simple truth is that anxiety’s foundation is based on not knowing something but as soon as you know, you have the ability to adjust and make a choice (I talk more about this in My Inspiration Never Dies).
There are therefore three types of knowledge:
- Known Known – The stuff we know we know
- Known Unknown – The stuff we know we don’t know
- Unknown Unknown – The stuff we don’t know we don’t know
Right now in the space of health technologies, applied technologies and AI we are NOT in the known known phase at all. In some cases we have very smart people who know that they don’t know but are trying hard to find answers but in most cases we, as consumers and governance makers are completely in the unknown unknown phase…we literally do not know what we don’t yet know (but many claim to know it all).
Where AI (Actual Intelligence) can help
In order to be able to really succeed and achieve any type of artificial intelligence we do need machines to learn but we also know that the biggest concern here needs three fundamental (human) questions answered:
- Why are they learning
- How are they learning
- What are they learning
The first has a common theme with a purpose to make a better, safer, healthier world for humans (we’d like to think). The second and third however get a lot more convoluted.
A computer needs data, the algorithm needs data and they learn by using the incoming data to process through a pre-determined thought pattern to then diagnose a solution or result. The issue is where a human designs the algorithm based on their life experiences or similar demographics and hands over the agility to learn from this seed for programmed autonomy. We have to remind ourselves that the human brain and its unique high speed processing capability of the vessel it controls (our body) and the environment it effects uses empathy and adjusts naturally so no algorithm can allow creativity and individualisation by default as it looks for patterns and habits and with all due respect, I am not the same as you.
The third question is where I have sincere concern, but a solution too.
The absolutely known known is that the data being used to support future artificial intelligence and to design our daily wellness programs is error ridden and this needs to change if we want to improve our lives today and significantly in the future as the population grows.
When does your watch take a break?
Claims that your watch can monitor you 24/7 for your life are false. There are fantastic outcomes in using wearable devices but the scales that tell you the lowest weight are not necessarily the best ones on the market.
My observations on daily applied technology have witnessed my wearable confirming I slept for 1 hour and 27 minutes while driving between Melbourne and Canberra. I have watched gym members scan themselves for body composition after a work-out wearing a rucksack. I have been confronted by weight loss challengers telling me that my $27,000 medical grade bioelectrical impedance analyser in a controlled test environment by a trained practitioner is useless as their $50 home scales tell them the right weight. I have seen athletes performing fitness assessments while being yelled at by practitioners in order to falsify a “better” result. I have witness talent selection processes being performed on wet grass then a dry court for two different groups and the data compared to select the best performers. I have watched sport coaches do high-load training sessions on players who are dysfunctional and just worked a 12-hour shift on a building site. I have seen two Physiotherapists perform completely different subjective assessments on the same patient during rehabilitation and use the results as evidence of further intervention needs. I have been asked by a personal trainer to adjust data on clients to show improvement when there was none. I have watched businesses profit from high scale, low quality assessment on school children to claim “well-being” improvement in physical literacy. I have heard from hundreds of PE teachers that they don’t even measure a 20m test properly and none of them know why they do a beep test on their students. This can go on and on (and I am available for talks 🙂
The reason I wanted to reiterate what my team and I see on a daily basis in our world of applied technology in assessing human performance is to highlight the following reason of why we need to take control and accountability today.
What if we didn’t change?
I am going to use one example that I am sure most people can relate to. A weight loss program.
Firstly, weight gain for many is as important as weight loss so we need to adjust our mindset that weight is only a component of healthy living and not the championship ring.
Standard weight loss program:
- Register to a gym where a qualified personal trainer will weigh you and take some manual measurements of your limbs and torso using a measuring tape. If a more proactive gym, they may use a bioelectrical impedance analyser but often will not weigh your clothes, ensure you haven’t worked out before the scan, have hydrated properly, do not have a pacemaker and are not pregnant.
- From the data a goal is set normally fixed on the weight (total mass), BMI (irrelevant score as you are not going to grow or shrink in height over 12 weeks) and maybe body fat percentage. For the real purist you may even get a Basal Metabolic Rating (BMR) too.
- A great report is printed telling you that you have an algorithm based metabolic age 7 years older than you are and the trainer compiles this data to design a program for you.
- You spend 12 weeks nailing it, working hard and aching but you feel great (this is due to the fact that actually doing any exercise is positive to mental health) but the trainer is motivating and the group are awesome to be involved with.
- At the 12 week mark it is time for the results and you are then re-measured possibly by a different trainer using equally unreliable and subjective limb measurements and very likely to perform the 2nd body composition assessment in different clothes and unlikely to have been similarly hydrated.
IMPORTANT NOTE – Before anyone reacts about the negativity towards the trainer here, I am not questioning their expertise and professionalism however under any level of qualification, a personal trainer has not been trained in validation of assessing a human and this is the point I am making.
The trainer designs the program for the client based on the initial baseline data that will rightfully assume is accurate, the clients’ goals and the capability they have in experience and resources.
If the baseline had just 3% error and the re-assessment had just 3% error the result can be 6% inaccurate therefore a 60kg client could not change at all yet be informed they had decreased by 3.6kg. 3% error is not hard to do. If you have a 35cm bicep, that is just 1cm between a tape that is too tight or too loose or just 1.8kg of clothing and water retention in a 60kg person when weighed.
Consider this when in the health and rehabilitation space. The difference between range of motion for a person recovering from ACL injury is essential for their overall life and being able to get back to proper movement.
The solution
Global Performance Testing Quality Assurance (GPTQA) was developed to simply provide support to the consumer and the professional in how to perform an assessment with reliability, validity and consistency and provide everyone with the confidence that the baseline and follow up assessments could be compared consistently (apple-to-apple).
The process has the capability of being integrated into future courses for student learning and in professional development for experienced health practitioners, trainers, coaches and educators that is agile to the innovations in technology.
The simplicity of the message is passed to the consumer so that they don’t judge the technology with negativity but embrace the known unknown (error) and take responsibility to learn themselves on what a good walk feels like rather than the number of steps accrued.
The firewall design of GPTQA ensures that current and future tech must not only meet technical and governance certification through controlled testing but that it can also replicate that reliability in the real-world. This ensures we have a clean reservoir of data to draw from allowing the architecture of artificial intelligence to be much more reliable and intuitive therefore creating confidence.
My thoughts
We need technology but we can’t be reliant exclusively on it. My observations over the last fifteen years have reinforced the dangers around misuse of apps, algorithms and technologies without the focus on the purpose and longevity.
If we are to grow to 8.5 billion people in just a decade, then we need to make sure the next 10 billion people who will rotate into this world over the next century are well informed and are not in the unknown unknown.
Smart investment in the future will be in clean data collection, interpretation and application for a positive habitual change for society and personal trainers of the future will be technologists.
Graham Dudley