11 Times AI Beat Humans at Their Own Game

Man developed AI to automate time-consuming tasks that need human intelligence. There is no doubt AI delivers what man designed it for. However, the question is whether it can surpass human intelligence. Elon Musk predicts AI will be smarter than the smartest human by 2026.  While we wait and watch, let’s look at 11 times AI outshined humans.

Deep Blue Wins Against Chess World Champion (1997)

Image Credits: Pavel Danilyuk from Pexels via canva.com

Deep Blue, IBM’s brainchild, became the first machine to beat Garry Kasparov, the world chess champion. The legendary player competed with Deep Blue in a six-game match over nine days. They played to several draws and a rematch before Garry Kasparov conceded defeat to the supercomputer. The machine could compute 100-200 million chess moves per second. 

Later, the player admitted, “I have to pay tribute. The computer is far stronger than anybody expected.”

Watson Trounces the Best Jeopardy Champions (2011)

Image Credits: Kaspars Grinvalds via Canva.com

Watson, the question-answering computer from IBM, won consecutive games in a Jeopardy challenge telecasted on television for three nights. The computer earned $77,147 against former Jeopardy champions Ken Jennings ($24,000) and Brad Rutter ($21,600). Watson could decipher the questions presented to it in natural language, thereby proving its intellect.

Ken Jennings had remarked in good humor, “Quiz show contestant may be the first job made redundant by Watson, but I’m sure it won’t be the last.”

AlphaGo Scores Victory Over Go Professionals (2015-2016)

Image Credits: Sora Shimazaki from Pexels via Canva.com

The world witnessed a historic stride in the technological realm when Google’s computer program AlphaGo outdid two well-known players in the Chinese board game Go. This game is quite complex and needs a strategic mind as it involves multiple combinations of moves. 

The first match happened in 2015 between AlphaGo and Fan Hui, a three-time European Champion. The computer program scored 5-0. AlphaGo created a buzz again when it scored 4-1 against 18 world title winner Lee Sedol in 2016. The computer program bagged 9-dan, the top rank for Go professional players. 

LipNet Reads Lips Faster Than Humans (2016)

Image Credits: View more by PhonlamaiPhoto’s Images via Canva.com

Oxford scientists created a software called LipNet using deep learning AI and a state-of-the-art speech model. This automatic lip-reading system can read lips with 93% success compared to 52% success of human lip-reading experts. It can predict words and complete sentences. 

Libratus Out-Bluffs Poker Players (2017)

Image Credits: Netfalls via canva.com

The poker-playing AI system Libratus astonished everyone with its superhuman intelligence. Developed by Carnegie Mellon University, this system pitted against four professional poker players in a 20-day competition. It leveraged computational game theory and reinforcement learning to beat these players in 120,000 hands. It led the game with $1,766,250 in chips.

This conquest of Libratus over humans was a noteworthy event as poker is the game of bluffs and the AI program did it like a pro.

AI Model Outperforms Humans at a Reading Test (2018)

Image Credits: PhonlamaiPhoto’s Images via canva.com

The retail giant Alibaba developed an advanced AI model with deep neural network technology. This model took the Standford Questioning Answering Dataset, a question-answer collection based on a set of Wikipedia articles. It topped the score at 82.44 as against 82.30 by its rival humans.

LawGeex Lawyers Up (2018)

Image Credits: Kittipong Jirasukhanont from PhonlamaiPhoto’s Images via Canva.com

LawGeex bot and the top 20 US corporate lawyers were asked to identify loopholes in five Non-Disclosure (NDA) agreements. The bot bested the human lawyers in both speed and efficiency.

The bot took merely 26 seconds to complete the work with a 94% accuracy rate. On the contrary, the human peers spent an average of 92 minutes and could achieve only 85% accuracy. 

OpenAI Team Outguns Video Game Players (2018-2019)

Image Credits: Michelangelo Buonarroti from Pexels, via Canva.com

DOTA 2 is a multiplayer competitive video game. It has several layers, strategies, and tactics challenging the players’ mental prowess and teamwork in real-time. OpenAI’s OpenAI Five, a group of five neural networks operated as a team to battle against five amateur human players in this game in 2018.

A year later, OpenAI Five made the headlines when it crushed OG, the world champion DOTA 2 team at the finals of a livestream esports game.

Bot Solves Rubik’s Cube at Unmatched Speed (2018)

Image Credits: View more by PhonlamaiPhoto’s Images via canva.com

Currently, Max Park holds the world record for solving a Rubik’s cube in 3.13 seconds. While his record remains unbeaten, it would be interesting to watch him compete with a bot created by Ben Kartz and Jared Di Carlo of MIT’s Biomimetics Lab.

This bot cracked the Rubik’s cube in merely 0.38 seconds in 21 moves. It uses cameras, actuators, and other motors to work its magic.

AI Robot Excels at Physical Skill Game (2023)

Image Credits: ArtificialArtist from pixabay via canva.com

AI has been surpassing humans in mind games so far. In a first, a robot called CyberRunner defeated its human counterparts in the Labyrinth game. Players are required to showcase their motor and reasoning skills to win the game. 

CyberRunner designed by Swiss University ETH Zurich has two motorized hands, a camera, and a computerized brain. It finished the game in 14.48 seconds, breaking the record of 15.41 seconds of former record holder Lars Goran Danielsson.

Faster & Accurate Healthcare Diagnosis

Image Credits: Kittipong Jirasukhanont from Phonlamai Photo’s Images via Canva.com

IBM’s Watson for Health and Google’s DeepMind Health are revolutionizing healthcare. They can accomplish some tasks quicker and more efficiently than humans. For example, Watson can store and review information from medical journals and case studies.

Some tests found that Watson had a successful diagnosis rate of 90% for lung cancer compared to 50% of human doctors. It also identified a rare form of leukemia in just 10 minutes. DeepMind’s AI can detect 50+ eye diseases as accurately as human eye professionals. 

16 Things Siri Can Do That You Didn’t Know

Mangostar Studio via Canva.com

The unexpected rise of Artificial Intelligence (AI) has transformed every sector, from retail to automobile, healthcare to banking, etc. Its impact is also being seen in the performance of stocks listed on varying exchanges. This article lists the 12 best AI stocks to help you capitalize on the AI trend and earn great returns.

16 Unexpected ChatGPT Capabilities You’ve Probably Never Tried

Image Credits: Gustavo Fring from Pexels via Canva.com

ChatGPT has 200 million weekly active users across the globe, 67.7 million in the US alone. You must have used ChatGPT for research, email writing, or social media content creation. However, ChatGPT can do some insane things that we bet you didn’t know. Read further to know.


Recommended