AI and coaching

As some of my readers know, my backup job in case consulting doesn't work out is to be a ski instructor (and yes I am PSIA certified). So most weekends in the winter I am at a local ski resort in the Lake Tahoe region of California teaching people how to ski. Sometimes it is an adult, sometimes a child. Sometimes a first timer and sometimes an advanced student and we are skiing fresh powder or learning to ski moguls. Skiing is a lot of fun, I recommend that you give it a try.

When the company Carv came out with an AI powered skiing coach I was of course intrigued - "Carv analyzes your pressure and motion 20 times a second to give you a Ski:IQ (TM) score on every run." It promised to provide feedback that would improve your skiing whether you were a beginner, intermediate, or advanced skier. "There goes my backup job!" was of course my first thought.

So I bought the device to see if it would really be a replacement for a human ski instructor and my experience with it is a great real world example of how AI complements and changes, but doesn't replace, what we humans are capable of doing, or at least not yet. As with other human work where AI is being introduced, human beings still have advantages in critical thinking, creativity, and interpersonal relations. The role of the human being though should change as it is augmented by the AI capabilities. I'll explain a bit more about my experience with Carv and let you draw your own conclusions for other work where AI can augment human activity and improve outcomes.

How Carv Works

The Carv system consists of four components:

1. A pad placed under the lining of each boot which detects pressure from different parts of the foot (toes, heel, etc.) as well as the rotation and edge angle of the boot.
2. A battery and communications unit which connects to the pad via a physical cable and transmits data wirelessly to a user's smartphone.
3. An application on the smartphone which in addition to collecting data from the pads also uses the phone's GPS to determine what resort (and even which run) you are on as well as speed and elevation change. The application has a voice mode that can provide realtime information and also coaching if turned on.
4. A web service that collects, stores, and analyzes the information, provides in-depth coaching tips (including videos), and also has a "leaderboard" for comparing your performance against other skiers on exactly the same runs.

As an example: I ran one of the "challenges" on the smartphone application called "Max Edge Angle." As I skied, after each turn I would hear the edge angle that I had achieved on that turn and then at the end of each run it would tell me the highest edge angle that I had achieved (66% by the way).

In the "coaching" mode I tried "Carving Training" and with each run I would receive a score and some suggestions about what I should be trying to work on to improve my score (achieved "Grade 12" and was told to work on my edge angle which is how I selected the Max Edge Angle challenge...)

How this Augments (not replaces)

The analysis and feedback from Carv has definitely helped improve my skiing. And despite the useful coaching tips (and videos etc.) there is still something that a human instructor can do that Carv cannot. There are several things I am doing when I am working with a student --

• Evaluating and overcoming the non-physical challenges they face when skiing - for example, many people experience fear when hurtling down a slope and this can interfere with their ability to learn much less use the right skills in any given situation
• Motion Analysis goes way beyond what is happening at the boot - a visual assessment of the full body position in motion is really needed to evaluate and recommend strategies for improvement
• There are many different learning styles -- some people are visual learners who need to see a demonstration, some are more analytical and want to understand the science, some are spatial, etc. Diagnosing how a student learns and giving them the right support is a key part of being successful.
• Especially teaching kids, having some creative ways of having them imagine what they need to do is important to get them engaged (it can help with adults also).
• Many different practice drills have been invented by instructors over the years and it isn't enough to just tell a student about a drill, but also to demonstrate and observe and offer corrections so that the student is practicing the drill in the correct way to learn the skills the drill is intended to improve.
• There are a whole bunch of basic safety issues in skiing (well summarized in the "Responsibility Code") which are really brought to life when observing what is happening around you on the ski slope.
• And an important role for an instructor is to also be a mountain guide, introducing (skill appropriate) terrain on a local mountain so that guests will get even more out of their visit.

My goal in this list was not to be comprehensive but to just highlight some of the things that the Carv sensor and application can't provide a student and where there are still opportunities for a human instructor to add value. What Carv can do though is to provide insight to the student about where they need to focus, and by sharing this with an instructor can complement the instructor's own visual motion analysis to help select the right areas to provide training.

In my view this is a great example of how we can be adopting AI across all of our human activities. While there are certainly areas where technology will replace human activity, any tasks that benefit from critical thinking, creativity, and interpersonal relations will still benefit from human contributions. What we shouldn't do (or be worried about) is where our machines can do a better job. A human observer will never be patient enough to watch 500 turns and accurately read out my edge angle and whether I am improving or not. But a human observer can help me focus on edge angle as a skill, give me a great drill to practice and tell me if I am doing the drill correctly so that I actually improve. And when I then do the next 500 turns, AI can tell me whether I am getting it right or not.

Where are the places in your work that a patient and tireless observer could help to give you the data and insight to help you do a better job? How are you thinking about using AI to transform your role in those tasks?