Amir Mirzaee, Chief Commercial Officer and Managing Director of Bayes Esports, dives into all things AI and its use cases within the wider esports industry – most importantly, how its misusage is hurting the esports industry
Artificial Intelligence, or just AI for short, has become a contentious topic in the recent past. Programs such as “ChatGPT” opened the door to possible uses of AI tools to the general public.
If given the right input of data, it can be used to create music or art, analyse theoretical or practical situations, or be the knowledge database of the future, constantly learning and adapting to what is happening around us. From entertainment to education, the use cases for AI seem endless in nearly every industry.
The esports data industry in particular is no stranger to the usage of AI. Generally speaking, there are two main use cases for AI models in esports.
AI and data collection: The fight against data scraping
More traditionally, AI in the esports data industry refers to the collection of match data in the first place. Match data scraping is a market practice that uses AI to scan the official broadcast of esports tournaments for all kinds of valuable information.
While mostly accurate enough to create post-game statistics and leaderboards, scraping data has proven itself to be too unreliable and slow for continued use in the esports betting industry in particular.
Since 2019, esports broadcasts are inherently delayed by at least 30-60 seconds to inhibit players from getting any outside information that may give away the other team. Any scraped data feed suffers from the same delay, which makes them unsuited to be used in the esports betting industry.
Furthermore, the act of scraping data bypasses the rights of tournament organisers and game developers to their data. By downloading the broadcasts to improve their AI for commercial purposes, data scrapers are violating the copyright of the rights holders.
With the increased usage of AI in various entertainment industries, we are seeing that data scraping is no longer just affecting the esports industry. Music labels in particular are now also pushing back against data scrapers that use the data of copyrighted songs to train their AIs.
These are then used to create new songs that imitate the sound of their source material, effectively replacing the original artists and putting their livelihoods at risk.
Whereas before AI and data scraping have been niche topics, we now find ourselves in a global and industry-spanning battle between two sides: the rights holders that try to protect their intellectual property on one, and the data scrapers that piggyback on the hype surrounding AI to promote their illegitimate products on the other. As a result, AI has become a controversial topic in esports that is strongly connected to the use case in which it is applied.
AI and data science: The future of the esports industry
The other main and more modern use case for AI in the esports data industry is the analysis of the vast amount of data that esports games produce. Due to the digital nature of esports, data of every action that every player makes exists in the form of official game server data.
Through AI, that data can be broken down in ways we may not have even thought of before, showing us exactly how the decisions the players take in the game impact the outcome of the match.
As an example, games such as League of Legends and Dota 2 present their players with the choice of over a hundred playable characters with their own unique playstyles. Every one of these characters can impact the outcome of a game in a different way.
This means that by choosing the right character to play, players can significantly alter their own winning chances before the game has even officially begun. Using AI and data science, that choice can be analysed, giving us a deeper understanding of how the interactions between different characters in various game situations impact the game.
Furthermore, AI can be used for integrity purposes. By having it analyse the behaviour of players during the game and comparing it to their previous performances, an AI can detect suspicious abnormalities in their play patterns. Similarly, it can detect irregular betting activities, allowing betting operators to take action against possible insider trading.
The industry is just now uncovering all that is possible with AI in esports and we can only imagine where that journey will take us. We have just scratched the surface.