AI Operating in the Background of the 2024 Paris Olympics
Operating in the background of the 2024 Paris Olympic Games was an array of Artificial Intelligence (“AI”) utilized by the International Olympic Committee (“IOC”) to pioneer ways in which the Games could be enhanced for both the athletes and viewers.
A key product of this effort was the IOC partnering with an AI-powered tool called Threat Matrix, designed to safeguard against online bullying towards Olympic athletes. Threat Matrix is a large language model (“LLM”) AI program which has the ability to recognize and interpret human language across extremely large data sets. LLMs can do so by drawing on a form of technology called “deep learning”, which enables it to understand how certain characters, words, or phrases function together to create text that reflects bullying or discrimination in over 35 different languages. The IOC hoped to use this AI program as a means of sorting through hundreds of thousands of posts each day, flagging those which may constitute bullying, and sending them to an immediate response team composed of human reviewers.
Threat Matrix was successfully piloted by the IOC during Olympic Esports week, where the AI successfully analyzed over 17,000 social media posts, flagging 199 to a review team who then determined 49 to constitute bullying. The program was then implemented during the Games, where it screened through endless social media posts across Facebook, Instagram, X, and TikTok, safeguarding the emotional well-being of the Olympic athletes.
Another innovative use of AI at the Games was through the IOC drawing on excess energy produced by local AI servers to heat the Olympic pools. The IOC did so by partnering with an American data company called Equinix, which operates a mass AI training site in Saint-Denis, France. The site generates large quantities of heat each day from its numerous servers tasked with training AI language models for well-known applications such as ChatGPT. Equinix was able to funnel the excess heat from its servers into pipes that were then sent to a local energy system used to heat the Olympic pools.
As AI continues to evolve, keep an eye out for how it may continue to pioneer positive change at future Olympic Games.
Author: Molly Podrebarac
Photo Credit: https://unsplash.com/@bozh_ntu
Expertise
Insights
-
Technology
TikTok Cleared to Continue Canadian Operations Following National Security Review
The Government of Canada (the “Government”) has concluded its most recent national security review of TikTok, the popular short-form video platform owned by Chinese technology company ByteDance Ltd… -
Technology
Confined Space Robotics Awarded $1.5M Contract to Automate Blast and Paint at Seaspan Shipyard
Confined Space Robotics (“CSR”) has been awarded a $1.5 million contract by Seaspan Shipyards to develop and integrate abrasive blast and paint robotic systems at Seaspan’s Vancouver shipyard. Seaspan… -
Technology
OpenAI and US Department of Defense Agree to Terms for AI Deployment in Military Settings
OpenAI has announced a deal with the United States Department of Defense (the “DoD”) that will allow the United States military to deploy advanced artificial intelligence systems within classified… -
Technology
Reinventing Cement: CURA’s Path to Reducing CO₂ Emissions
Global average atmospheric carbon dioxide (“CO₂”) concentrations have climbed roughly 25 percent since 1980. This is largely driven by fossil fuel combustion and intensifying wildfires, as the… -
Technology
Carbon Robotics’ Large Plant Model Brings Real-Time Plant ID to Laser Weeding
Carbon Robotics, the Seattle startup behind the LaserWeeder – a robot fleet that uses lasers to eliminate weeds – has introduced a new AI system called the Large Plant Model (“LPM”). The company says… -
Technology
Etsy to sell Depop, Gen Z’s Favourite Hub for Vintage Clothing, to eBay This Year
According to a recent article from TechCrunch, virtual retailer Etsy is selling Depop, Gen Z’s preferred online platform for buying and selling vintage clothing, to eBay for US$1.2 billion in…