AI Comes to the Olympics
The Olympics have come a long way in the nearly 120 years since the modern Games first reformulated the Ancient Greek tradition. I learned this week, for example, that for the first half of the 20th century, the Olympics awarded medals for competitions in the arts including sculpture, literature, and painting, in addition to the traditional sports competitions. I also learned that tug-of-war was an Olympics sport from 1900 to 1920 with the United States sweeping the podium in 1904! Who knew.
The evolution of the Games this year brings many of the typical changes that come from cycle to cycle, including the introduction of a new event. This year’s new kid on the block is the sport of breaking (i.e. breakdancing).
But the more meaningful changes are ones that mirror the rapid inroads that artificial intelligence has made in our lives. AI is now a part of the Olympics. Some of these changes may help to improve the Games for athletes, attendees, and fans alike. Some are downright creepy, and portend the beginning of a new era, the likes of which George Orwell himself would marvel at.
Here are some of the highlights of how AI will play a role at this year’s Olympic Games in Paris:
Al Michael’s voice: the iconic sportscaster known for, among other things, his famous, “Do you believe in Miracles? YES!” call in the 1980 Olympics semi-final hockey game between the United States and Russia, will not be at the Olympics this year. His voice will be, though. NBC has trained artificial intelligence on thousands of hours of Al Michael’s voice, which will now be used to recap highlights on Peacock. Per Vanity Fair, Michaels was originally skeptical, but once he heard examples of the AI, he came around, saying, “Frankly, it was astonishing. It was amazing. And it was a little bit frightening.” No word on how much Michaels might be earning from NBC for the rights to his voice.
Judging: sports like gymnastics and diving have a high degree of nuance and technicality to them, featuring elements and movements that even the most highly trained human eyes can not pick up on perfectly well. Plus, judges are human (at least up until now), and can be swayed by a whole slew of non-performance variables like the energy in the gym, their own mood that day, or the country of origin of the performing athlete. These are among the reasons why the International Gymnastics Federation has been testing a device known as the Judging Support System (JSS) over the past two years, which they believe can bring neutral, non-biased, and technologically sound analysis to the judges’ panel. The JSS will be used at the Paris Games not to replace human judges, but to bolster the human element and perhaps offer a check-and-balance. You can read more about this in a piece in the MIT Technology Review by Jessica Taylor Price.
Athlete resources: Intel Labs along with a Paris-based company called Mistral AI, has created AthleteGPT, which is a large language model AI tool for athletes specifically for them to be able to questions of the device on site that specifically pertain to them, like how do I find my venue, where is the nearest cafeteria, how do I find transportation, etc.
Athlete performance: Sumeet Kulkarni of Nature details several ways AI is helping to boost athlete performance including another Intel device that tracks 21 different data points on an athlete’s body that can be analyzed in both training and competition to maximize performance. Kulkarni also details how AI is being used to specifically design custom shoes made to the exact specifications of an athlete’s foot and to optimally create a nutrition plan leading up to events.
Athlete development: the IOC has partnered with Intel another way: to help find and develop athletes in under-resourced areas. Together the IOC and Intel launched a program in Senegal and tested 1,000 young people to see who might be the most likely based on a variety of metrics to have success as an high-level athlete. They identified 48 individuals who were not previously on an athletic development track who may now get enhanced opportunities to pursue an athletic career or even their own Olympics dreams.
The organizers of the Paris Games are promoting the use of AI in ways that are intended to make the fan experience more fulfilling, too. The wealth of stats available to fans, for example, who are watching broadcasts or who are actually there in-person is staggering. Much of the statistical analysis is available in real time.
The AI fan feature I am most curious about? There will be opportunities for fans attending the games to participate in a test that would show them the Olympic sports they are psychically best suited for. As it turns out, my college roommate is a near-Olympic level marathon runner. He has run several marathons in the 2:20 range, and even ran a 2:19:01 at the Toronto Marathon. He is attending the Games in Paris as a spectator and told me he would try to find one of these stations so they can tell him what sport he has the most potential in. My guess? It’s probably not breakdancing.
What Comes Next
The ways AI is being rolled out at the Olympics this year does feel a bit like a through the looking glass moment. It feels like an inflection point. If a television production company like NBC, for example, can create a large language learning model AI tool to recap highlights and use Al Michaels’s voice for it, what is the role of the future broadcaster? Will viewers push back on AI-narrated recaps? Or will production companies eventually save millions of dollars by “employing” chatbots as broadcasters instead of human beings.
The story is the same with judges. Any sports fan out there knows there are countless examples of umpires, referees, and judges getting it just plain wrong in major moments. But do we really want the balls and strikes in a baseball game, for example, to be called entirely by video umps? The technology is possible now. Minor League Baseball has tested it here in the United States. You can see computerized decisions at tennis matches, for example, when a video replay of a shot zooms in on the smallest detail of whether the ball touched any piece of the white line, and it can do it almost instantaneously.
But what is lost by not having the human component, especially in a sport as nuanced as baseball or many of the other Olympic events (and actually, baseball is sadly no longer an Olympic sport…). That is a question not just for the Olympics and not just about sports, but for all of society as AI is inserted into more and more of the traditional work that human beings have always done around the globe.
Ben Sprague lives and works in Bangor, Maine as a Senior V.P./Commercial Lending Officer for Damariscotta-based First National Bank. He previously worked as an investment advisor and graduated from Harvard University in 2006. Ben can be reached at ben.sprague@thefirst.com or bsprague1@gmail.com.
Weekly Round-Up
Here are a few things that caught my eye this week:
Friday’s Personal Consumption Expenditures (PCE) report showed another month of slowing consumer spending. I wrote about this a couple of weeks ago: things are not deteriorating, but they are definitely starting to roll over a little bit as far as spending is concerned. We have now seem year-over-year PCE spending go from 2.7% in April to 2.6% in May to 2.5% in June (which was the report released on Friday). The news on more dour spending was met cheerily by Wall Street, however, as the slow down may give cause for the Federal Reserve to lower interest rates this fall, which investors feel would be a positive thing for the business sector and therefore the stock market. The Dow was up 654 points on Friday with all of the major indexes up more than 1.0% following the morning release of the PCE report.
The National Multifamily Housing Council released a report last week indicating they are seeing loosening conditions in the rental market. A couple of the highlights, none of which will be a surprise to long-time readers of The Sunday Morning Post:
A lot of new units are coming online, which is easing some of the pressure in the rental market
Demand for rental units remains strong, which has led to all of this new supply being absorbed into the market in a healthy way with some sense of equilibrium.
After several quarters in a row where financing conditions tightened and the number of deals contracted, the last two quarters have seen some increased willingness by banks and others to finance transactions, which has led to better deal flow.
Plans by Silicon Valley investors to build a brand new city of up to 400,000 people near San Francisco are now on hold. Local opposition, environmental impact, the vagueness of the financing plans, etc: you name it, a plan to build a new city is obviously going to have a lot of hang up’s.
Have a great week, everybody!
Very interesting piece Ben. AI AL is fascinating. I feel like in the future, where we already have all of these alternate broadcasts, (Manning-cast, etc), we will be able to select our broadcasts like we do our soft drinks at Burger King. Giants-Cowboys, I think I feel like some Madden & Summerall tonight and AI can generate that experience.