Artificial Intelligence from the context of games selection.
Artificial Intelligence from the context of games selection.
Over the last few months I have been meeting with various CEOs, CTOs, and some of the world’s most prevalent, forward thinking AI groups, companies and societies from around the world. My travels have taken me to San Francisco, Helsinki and London. My hope with this article is to share my perspective, which hopefully proves to be illuminating, on the subject of Artificial Intelligence.
What is Artificial Intelligence?
“The simple answer is that a machine that exhibits intelligence, in its broader sense AI, is a powerful tool which helps us both make sense, and put to good use, the vast amounts of data which modern technology provides. Implemented by employing mathematical models and algorithms, that enable the extraction of knowledge out of a large amount of data, these models and algorithms are often designed to imitate the same learning principles as the human brain” – ANDi Games Ltd Chief AI Officer, Babak Takand
How is it created?
Creating Artificial Intelligence which can match or surpass ours is still unclear. The ability to learn and apply knowledge is the subject of Machine Learning, which consists of computational models and algorithms. These computational models can be found with most modern programming languages including; Python, Java, C++,C and Lisp.
But why now?
Artificial Intelligence has been in development for over 40 years and is progressing continuously, especially because as hardware improves, so does the capability of the software. Last year there was discussion around Big Data, which is one of the most important aspects of AI. It’s the collection of large quantities of data, and in turn, how we make sense of that data to allow technology to make decisions and take actions.
The part I find most exciting is that we continue to learn more about ourselves and automate what software does to benefit us. Let’s look at Artificial Intelligence in games to explain more;
The first real software intelligence was developed by Arthur Samuel who created the world’s first self-learning program, to play Chess. Having fathered the term “Machine Learning”, he created the Samuel Checkers-playing program on an IBM Supercomputer with powerful memory, processors, and software that could calculate over 300 million decisions per second and identify the optimum move. He later went on to defeat the world champion Garry Kasparov in 1996.
The next breakthrough was AlphaGo, developed by Google’s DeepMind in London. They championed the world’s best players of the classic Chinese game Go against Lee Sedol, who holds 18 major Go titles. This all happened just last year! This uses a different technique called Deep Neural Networks to identify patterns, analyzing different categories through 48 layers of computer neurons; this essentially mimics the idea of intuition and creativity.
AI in video games
However, it’s not only with board games that the software has seen continued success.
Kris Graft provides a fantastic overview of AI in video games, explaining;
“we think of AI in terms of obvious AI agents like NPCs (Non-player Characters) or enemies that duck and roll away from gunfire”.
Over time the NPCs improve and the game can act on its own. So what does this mean for the player experience? Improved AI may mean the machine becomes too experienced to beat, in other words, so good at its function we cannot control it.
CEOs, investors, philosophers, publishers and even us, are looking forward to the future of computer consciousness. How can we as human beings manage and predict its improvement, safely test or even determine what the different levels of a “conscious computer mind” can be? And what it will create from its own will? No one knows what mistakes, decisions, findings, or characteristics the ultimate machine will have, simply because it does not yet exist artificially.
ANDi focuses on recommendations and prediction
“Recommendations“ we have seen, but “prediction“ is best explained by what Amazon are doing with Alexa.
“Imagine you’re about to leave the house to pick up your kids. As you grab your keys, you hear a voice from the device on your coffee table: the fridge analysed its contents “It looks like you’ll use the last of your milk tomorrow, and yogurt is on sale for $1.19. Would you like to pick up an order from Trader Joe’s, for a total of $5.35?” You say yes, and Alexa confirms. The order will be ready for curbside pickup, on the way home from your kids’ school, in 15 minutes”. – Harvard Review
So, we look at how the market recommends mobile games currently. They analyze the previous genre you play. For example, you play a stick game for 3 minutes, you should try another stick game.
ANDi however, broadens the data analyzed around the subject, and by applying machine learning ANDi identifies trends in user habits, correlating them with games played on mobile, or tablet, and determining individual interests. Additionally, the average user has more than one mobile game to play at certain period of the day or month.
An example of ANDi in action would include; “You are about to leave the house in 10 minutes for work, ANDi sends you the following recommendation; ‘Quickly download this particular brain teaser that you can play and enjoy offline while on the train, activate your mind and prepare for the working day!’“
Once we can single out the appropriate application from millions of options, we tailor the notification timing for people at moments that best suit them.
The user comes first.
Other applications for Artificial Intelligence.
AI is already being used in a number of areas and there’s more developing all the time. From improving health, medicine, innovation, energy, environment, manufacturing, and even general tasks, we can clearly see how this will affect the larger population.
However, as Artificial Intelligence is used, what we can say is it will reflect society’s needs. We look forward to seeing the progression of AI in 2017!