A blog where you can read a myriad of articles from diverse fields such as economy, psychology, philosophy, science, history etc.

Saturday, 18 June 2016

Artificial Intelligence

Most of you are aware that artificial intelligence (AI) is the intelligence of machine in performing a task that was once conceived to be only within the reach of a human brain. But this definition of AI evolves over time as technology spreads and some tasks become common. So to put it concisely, AI is whatever haven't been completely achieved yet. Human started building machines initially to help them in their physical tasks but in the past 100 years, humans have embarked on a journey to create technology to enhance their brain and this led to an artificial intelligence race among many top companies. Companies like IBM, Google are pouring hundreds of million dollars in AI technology. Before discussing the role of AI and what it means for the future of human race, let's briefly allude to the history of AI.

Brief History of AI:

In the 1940s, Turing's theory of computation suggested that the machine by just shuffling 0 and 1 can represent any conceivable physical laws of  this universe. In 1956, John McCarthy and others started writing programs that helped the machines to solve problems in algebra, playing games with human and speaking English. This led to heavy funding of AI related research by US Department of Defense. But they had overestimated the capabilities of AI and hadn't factored the issues such as lack of computing power and lack of sufficient data. This led to a fall in the AI related research. But as the computing power grew exponentially as forestalled by Moore and the advent of big data and the internet and cloud meant that there were a lot of data to be processed and also the computing efficiency was now faster and cheaper and also these data were easily available to any internet connected system.This was pristine milieu for the rise of AI.

Image Courtesy: map.norsecorp.com

Significance of AI:

So how is AI useful, in this ever changing world one needs to keep track of all the changes in this world to be considered as an expert in a field. Say 200 years back, one could have specialized in 10 fields due to limited availability of data. But with the big data technology, there are huge data being generated per day which is beyond the capabilities of any human to process. Thus nowadays, we see people specialize in a sub-section of any field. AI, thus, help people in being abreast of latest trends and information about the desired field. An example of such AI would be an expert system which can help a doctor treat its patient by recognizing the symptoms. Another example would be when an AI help an e-commerce company to market its product to a specific target group or even individual  by identifying the specific target group or individual. These AIs is becoming more and more intelligent at their assigned tasks.  In 1997, Deep Blue,an AI developed by IBM, defeated Garry Kasparov ,who is considered to be greatest chess player of all time, by 3.5-2.5 . Google's AI AlphaGo and Deepmind have defeated many legendary Go players. Expert systems are helping many companies to increase the sales of their product drastically. So what does the future holds in store with the growth of computing power and amount of data for the machines to learn?

Is future with AI a dream or a nightmare for human race:

What will happen once we develop a real AI, in the sense that it can learn and grow itself,whenever there is a change in the situation, it can change itself to better fit. Also, when all the different AIs can collaborate together to achieve certain tasks.What will happen when these super intelligent machines start competing with 7 billion people for resources to achieve a certain goal. Let me present a hypothesis:

Suppose an artificial intelligent system is created with a task to maximize the production of diesel cars, now to produce these cars it will need all the raw materials. Now there are a lot of practical issues like trade barrier and environmental concerns that need to be accounted for. In an ideal scenario, it would want to take control of the entire world to have indefinite resources available to produce these cars and it can turn against humans whose values doesn't align with its values.

Also, with these AI  residing on clouds and distributed across various systems around the world, it wouldn't be easy to shut them down. Also, if we are able to build a single kill switch, these machines being super-intelligent would find a way to protect it.

Recognizing these potential threats, many AI researchers have started building machine ethics which teach these AIs about human ethics and values so that the control remains with humans.






Upcoming Post

An attempt to understand the human brain and the nervous system

From eating, breathing to thinking, analyzing various scenarios and feeling, everything is controlled by your brain. This jelly-like mass ...


A blog where you can read a myriad of articles from diverse fields such as economy, psychology, philosophy, science, history etc.

Powered by Blogger.