After almost a decade of wearballs testing, I have really collected a terrible amount of health and fitness data. And when I enjoy poring on my daily data, there is a part that I have come to hate: AI summary.
In the last two years, a holocaust of AI-borne summits has been sprayed in every fitness, welfare and wearable app. Strawa introduced a feature called Athlete Intelligence, carrying your raw workout data as AI and relaying it to you in “plain english”. What is the WHOOP coach in WHOOP, an AI chatbot that gives you a “daily approach” reports, in which the weather, your recent activity and recovery matrix and workout suggestions are summarized. Ora added OURA consultant, another chatbot that summarizes data and draws long -term trends. Even my Bed Every morning congratulates me with summary how his AI helped me sleep every night.
Each platform has the nuances of AI, but the summary of the morning goes a little like this:
Good morning! You slept 7 hours last night with a relaxing heart rate of 60 bpm. This corresponds to your weekly average, but your slight height heart rate suggests that you cannot heal completely. If you feel tired, try to go to bed first tonight. Health is all about balance!
That might be seem The accessories, but those summings are usually placed next to a chart with the same data. This is worse for workouts. Here is one that Straw’s athlete intelligence has recently generated for a run:
Acute runs with high heart rate areas, pushing into anaerobic field and well logging over a relative effort above their specific range.
Thank you? I can ask athlete intelligence to say “peacock”, but it rebuilt my effort, heart rate area and pace metrics that I can see in the graph in the workout summary. If you did not know anything about my athletic history or the circumstances around this run, then this summary could read as practical. What is summary here:
- Trying my mileage in only my second run of the year was dangerous, high humidity, 85-degree-plus weather, and the history of my spotted workout in the last two months in the last two months as compared to its six months. Street season has access to data And I have done every workout in the last five years.
- I had to shorten this run because I fell and cut my hands and knees. It is known that Strava has access, as I have uploaded a gramly picture in addition to text notes. After adding the above notes, the updated summary only indicates that I have shortened the run. My injury did not change anything about its insight, even though this is the most important thing that happened during this run.
A more useful insight can be: “You run away during the record-breaking summer for your area. When you maintain a consistent and stable speed, you have a bad habit of taking advantage very quickly after a long break, which is a pioneer for many self-reported injuries in the last five years. The wounds are cured.”
Ranna, a popular running app that includes AI Insights, one was born Little More useful summary. It said that my next run should be “easy”, which is completely on time to recharge for me. I am sorry, but 48 hours is not enough time to safely heal my knees without the risk of reopening my wounds.
In-app chatbots are not much better. Yesterday morning, I asked the whip coach if I should run today because I had injured myself on my last run. This told me: “It is unable to respond to the message you sent. Please try to send a different message.” I again told my prompt, “I am injured and a lame. During recovering, generate a low intensity workout option.” I was asked to contact the What membership services to continue the conversation.
The OURA advisor was more helpful, paying attention to my daily summary: “With your readiness and recently with stresses such as heat, an injury and high glucose, your body can feel more tired than normal today.” This suggested that I prioritize comfort. When asked, “What kind of movements are fine when the injured knee and a minor lame?” If there is no pain, light stretching, and a reminder to fully relax, it responded like a small and easy walk with the answers of general knowledge. This is close to an ideal response, but I had to guide it for the type of answer I wanted. Are insight so It is a general-purpose that they benefit self-convenience newbies-and still, only if they are allergic to Googling.
My bottled run is actually the type of landscape where Tech CEO says AI Insights can be the most useful. In theory, I agree! it Would like It is good for being a competent, underlying chatbot that I can ask more fine questions.
1,3
For example, I have done an irregular sleep program this month. I asked the OURA advisor if my sleep and readiness trends have shown signs of a high risk of injury. I also asked if I have a high level of unusually sleeping debt this month. In both cases, it was said – it was said that I was Improvement in,
As a result, there was an hour -long debate with a chatbot, which left me questioning the experience of my own life. When I asked to fly in a particularly stressful week earlier this month, it told me that its insight was “limited to the most recent week and current trends.” Such a six -year value defeats the value of the OURA data.
After months of refusing Reddit and other community forums, I Know I am not the only person who finds these AI facilities to laugh. And yet, OULA chief product officer Holi Shelton told me that OUVA consultant’s response has been “highly positive”, with 60 percent of users it is used several times a week and 20 percent daily. “Beyond the frequency,” shelton says, “it is providing real effects: 60 percent of the advisor says that the advisor has helped them to understand the matrix or health concepts in a better way that they were found to be confusing earlier.”
Meanwhile, Straw spokesman Brian Bell told me that the athlete intelligence was to help the initial athletes and “the feature response remains strong” about “80 percent with those who react with those who react to ‘helpful’ to respond to ’80 percent of the features’ to find the feature for ‘helpful’.
A Whoop spokesperson was not able to respond by publication.
This milvetost summary? They are probably the best compromises between speed, cost, utility, data privacy and legal liability
I understand the stem of my frustrations by the underlying boundaries of the LLM and the disturbances of private health data. Straw can be a real fitness data hub, but it is lacking All Health data points required to create overall, useful and individual insights. The OURA advisor will take a long time to crunch a year’s sleep data for trends. That delay is guaranteed to offer a poor user experience. To not mention, they will be prone to their membership from $ 5.99 per month to add that type of computing power. I am not sure, but the whoop coach has rejected questions related to my injury to protect myself from liabilities when something bad happens to me. This milvetost summary is perhaps the best agreement between speed, cost, utility, data privacy and legal liability. But if this is the case, let’s be honest. Current AI features are rebuilding data, much such as the book report is written by a fourth-gradder, which is dependent on the Wikipedia summary rather than reading the book. It is a feature with duct tape and a dream because AI is zegotist. Perhaps One Days, these AI insight will create a useful and personal experience with an insightable insight. That day is not today, and it is not worth paying extra.
