We are currently in the 21st century where people are almost entirely dependent on machines. In which technology paved the way of major changes in our lives. Applied science is rapidly making dreams into a concrete vision of reality. It is merely because we are in the generation of full of incremental advancement and evolution. Elon Musk is one of those techies behind these innovations. He is a South African-born Canadian, American entrepreneur, a multimillionaire and a quest for an imaginative future. A brilliant engineer who founded Paypal, the one that developed solar-power systems, electric cars and a futurist who has a plan of creating a private space travel and even building a human colony on Mars. He introduced a machine called Artificial Intelligence that reacts and executes as a human but still behaves irrationally. Artificial Intelligence might be useful in any medical diagnosis, fraud, and robotic ascendancy.
One fascinating issue emerged was when AI defeated world’s best video player in one-on-one match for Dota 2 game. It is an internet robot which interacts with human messages and commands through online. This learning machine already had beaten many pro players of Dota 2. Dota stands for Defense of the Ancient which is a multiplayer online battle arena (MOBA) computer game. A player chooses a unique character of different abilities known as “Hero” and has the control to make the hero move from any direction, attack the opponent and release weapons. This game is complicated because you need to protect your base and yourself while striking your enemies just to take the throne, else you’ll get killed by the opponent.
Introducing Go as a complex strategy board game, and chess, which is also far from being an easy game, Dota 2 was considered as more complicated than Go and Chess.
There are 113 heroes in Dota 2 that are divided into 3 categories and a lot of different items to be purchased to level up your hero. Is it really possible for a learning machine to function and defeat a real player for this human designed video game?
A lot of international competitions are being held for Dota 2. In fact, it is the world’s richest e-sport which cost millions as the prize pool for the winner. It was designed by Icefrog and was developed by Valve Corporation. Every year, Valve anchors annual tournaments and releases a prize pool that may exceed to $24 million championship prize pool.
But one headline brought astonishment to everyone when Danylo “Dendi” Ishutin, a Ukrainian professional Dota 2 player who is playing for Natus Vincere since 2010 was defeated by a bot. With Dendi’s 7434 MMR, he is the world’s 44th highest earning player as he reached $730,000 USD at 2014. He started his career at the young age of 17 and started playing professionally with WG in 2006. While the bot was only trained for 2 weeks and was able to store the game procedures and functions of heroes within that limited time span. During their live one-on-one match, Dendi expressed that the bot feels like a human, but a little like something else. After being defeated twice in their live one-on-one match, Dendi surrendered to the 3rd match. Dendi labeled the game as fun and challenging, he also said that he might need more practice to beat the bot. It was considered as an achievement compared to those competitions of AI Bot versus human for Jeopardy and Go game. It is because of the complexity of the Dota 2 game consisting lot of heroes and items to be manipulated.
It is actually not the first time in history to develop a bot player for online game, Microsoft developed an AI before intended for the famous 1980’s video game, Ms. Pac-man, in which human players that are not adept can’t beat the bot.
Since Dota 2 is a game of two teams consists of 5 players each team, the OpenAI team has a plan of making a 5-vs-5 match game to be possible by next year’s tournament.
Elon Musk explained that though AI might be useful, beneficial and gives amazing impact to everybody, those governments wishing to have Artificial Intelligence to be implemented in their country should be mandated by rules and regulation for the use of AI. Because it may still cause risk and danger for whatever purpose is if it will be used improperly. Meanwhile, OpenAI announced that they are releasing their bot for practice to all players who want to beat and play against the bot. Making their bots available for practice is just a give and take concept, in which the organization is making a new way to improve the bot through trial and error while gearing the players towards their improvement.
You can watch the match here :