The 2020 Olympics featured a new player, one that is likely to occur at every opening ceremony in the foreseeable future — AI.
The Tokyo Olympics of 2020 were the first where AI was a prevalent part of the games. Everything from robots, wearable sensors and machine learning were employed to analyze the massive amount of data produced from the athletes in real time, according to Deloitte.
At the 2024 Summer Olympics, AI is playing an even bigger role. The International Olympic Committee president has overseen the third AI strategy plan. In it, efforts like cybersecurity protection are documented, along with the plan to use AI to monitor energy consumption during the 2024 games and work with Intel to create digital twins to plan for future games. Not to mention the AI that the competing athletes will wear during their events, like the oldest men’s basketball player to ever step on an Olympic court, LeBron James.
James is one of many athletes at the 2024 Olympics who use the wearable tech WHOOP to track and analyze their health and movements while competing.
With AI playing this big of a role in its athletes’ performance, how do WHOOP’s own people employ the technology? Built In Boston spoke with Hilary Gridley, Director of Product Management at WHOOP, to hear how she uses AI in her day-to-day work.
WHOOP is a fitness tech company that created a wearable device to empower members to perform at a higher level through a deeper understanding of their bodies and daily lives.
How does AI help you in your day-to-day life?
AI helps me in a wide variety of ways. When I am working on visuals, I often ask ChatGPT for ideas for in-app visuals like icons and illustrations, then ask for examples that I can send to an illustrator. This helps me come up with better ideas than I could on my own, and communicate them visually. We also have an amazing content designer who has set up an army of GPT “interns” who can write UX copy.
When using AI to research, I can create extremely detailed opportunity maps in less than an hour.
AI in Action
Gridely is among the minority of adults in the United States who use AI, which a Pew Research Center survey found to be at 23 percent in February of 2024. However, Gridley is among the majority of business leaders — 64 percent — who believe that AI can improve productivity, as noted by Forbes Advisor.
Some of the top ways Gridely uses AI are:
- Ideating visuals
- Writing copy
- Conducting research
- Reducing busywork
What influence does AI have on your work?
AI helps me move faster on all sorts of tedious tasks. Yesterday, I was working on our Board slides and needed to incorporate some graphs that were all formatted differently. Rather than rebuild each of the graphs in the correct colors and styles, I simply shared a screenshot with ChatGPT and asked it to reformat the graphs using my specific instructions.
After a bit of back-and-forth, I was able to simply copy and paste the graphs into my slides, and the whole thing looked polished. This saved me a bunch of time, which is always extra appreciated when I’m on a deadline.
Can you share an example of AI helping you to create an opportunity map?
I recently wanted to understand what types of people with chronic conditions might be well suited to wear WHOOP. I mapped out a classification of all diseases, the most prevalent conditions in each category, the prevalence of each of those conditions in the United States and abroad, associated symptoms, risk factors and associated healthcare spending for hundreds of conditions. This allowed me to quickly size up the different opportunities and identify the most promising ones.
A task like this would have taken me a week if I didn’t have ChatGPT to assist. Instead it took me about an hour.
AI can pose some challenges too. What are some that you have faced and how did you overcome them?
The probabilistic nature of large language models can make it difficult to know how much trust you should put in what it tells you, especially for questions where it is more likely to be pulling from outdated or incorrect information. This has made it challenging to rely on its reporting on analytics or internal knowledge.
We’ve found that teaching the LLMs to cite sources so you can double check its information, or simply building in a human quality assurance test for any high-stakes decisions, has helped with this issue.
