Artificial Intelligence and Strength Training: Part 1-The Impact of Information
Current news is awash with stories about how artificial intelligence (AI) will shortly penetrate every aspect of our lives.
Depending on the source, predictions are being tossed around about AIs potential to eliminate zillions of jobs in areas where expertise is the thing of value.
Having worked in AI off and on since the 1980s, I tend to be skeptical of many of the forecasts tossed around.
In this article, (and next week) Ill discuss what I believe are potential impacts of AI in strength training.
What part of the problem is information?
If you are trying to improve your strength, the first question to ask is what part of the problem of maximizing my strength performance is dependent on information?
Information is only one aspect of improving.
There are physical performance requirements and developing physical skill.
This requires doing physical work AND doing the work (practicing) in a manner that enables you develop your physical skills to the greatest degree possible.
The starting point is to have information.
If you have poor information and thus do a crappy program you are not going to make much progress.
Thus, the first thing that is necessary is good information on what to do.
Here is an example that should illustrate that point.
One of the greatest sprinters ever to set foot on a running track was Usain Bolt. He had other worldly speed and was virtually unbeatable in his prime.
His complete training program was available in books, articles and on video.
Tens of thousands of sprinters around the world followed his program.
Many probably realized more of their potential than they would have with a lesser program.
What is unknown is whether the thousands of other runners knew how to hone their physical skills such as stride length, body positioning, arm movement, foot strike, etc. to realize their full potential.
Without guidance or excellent coaching on the minute details of performance, athletes, musicians, surgeons and anyone else doing a job that requires both mental and physical skill, the person stays at a mediocre (or lower) level of performance indefinitely.
The point is that simply knowing what to do is NECESSARY but not sufficient by itself.
Information is only the starting point on the long and difficult journey to mastery.
An Experiment
Now that I have made a bit of noise about how inadequate an AI program may be for building ones athletic skills, I decided to see what type of training package I could put together using Chat GPT.
I first asked it to prepare a training program for a 60-year-old experienced powerlifter.
The program I got sounded like something out of the standard supermarket magazine. They did build it around squat, bench press and deadlift, but used a 5 day per week schedule, including a few low value exercises, and generic advice such as 8 hours of sleep and drink lots of water.
The program recommended repetition ranges of 3-5 on all exercises.
I think you can see where this is going. The program was basically for beginners with a few nods to standard powerlifting approaches.
I asked about the proper technique for each of the power lifts.
The results were again dismal.
But Doesnt AI Know EVERYHING?
IMHO the reason the programs that Chat GPT built for me were so lame is that they were contrived using data from the internet. This skews the information heavily toward those who make YouTube videos and/or put up a lot of information on websites.
Unfortunately, some of the greatest power coaches do not publish or do vids.
In building the AI program it appears that the large data model used does not distinguish between advice given by an elite coach or advice offered by just another gym bro.
We get the average of everyones views.
Im reminded of comedian Alan Shermans song about good advice.Its all for free and worth the price
Next week Ill dig deeper into what you may or may not be able to get from AI in your personal quest to lift bigger ironor bumper plates.
Lift Big!
Richard
0 Comments