Artificial Intelligence is the greatest invention of the 19th century. With the earlier inventions of weak AIs to the more advanced and strong AIs, the technology has always proved to be useful. But it also sparks some fear. We have all seen those dystopian sci-fi movies where machines take over human beings, so the concern of some people is understandable when a robot is so perfectly imitating a human being that the only difference is its appearance.
To understand how an AI works and what are some of the evident pros and cons of AI implementations in human life, we first have to understand what an AI is?
Artificial Intelligence
The term Artificial Intelligence was first coined by John McCarthy in his academic conference in 1955. Even though the invention of AI is older than that but this period is considered to be the birth of AI. Before that AI was only a dream, a work in process. McCarthy is also considered the father of AI. Since the birth of AI, the technology has gone through some busts and booms of popularity. A new invention or discovery would spark a sudden interest of common people and everyone would be talking about it and after a while, nothing. But the search goes on.
In simple terms, the process of simulating human intelligence by machines is known as Artificial intelligence (AI). These processes include learning, reasoning, and self-correction. Depending on the abilities of an AI program, it is generally categorized as weak or strong AI. An example of weak AI is Siri, Google Assistant or similar interactive programs limited to work on a data collecting and analyzing basis. They are designed to assist human beings under certain rules, they will collect the data and use it for your assistance, like a simple data storage system of your mobile phone that stores your data for a temporary period to enhance your user experience. A strong AI is when a machine is capable of solving problems without human interaction by learning from its past experiences.
Various Types of AI That Has Changed Human Lives
From 1956 to 1982, the enthusiasm for AI-led to the discovery of different subcategories of artificial intelligence. Many subtypes emerged from the same root with the same vision. The hype it gained by McCarthy’s academic conference, inspired many tech enthusiasts and dreamers to build the future. AI has since been divided into some of the generalized categories including:
Automation AI
These are rules-based programs working on a strict, if-then-else basis. Robotic process automation (RPA) is when a machine can adapt to changing circumstances, it also falls under the category of automation AI. Sometimes automation AI is as plain and straightforward as a packaging machine that is programmed to repeat the same task in the same way over and over again. While sometimes it can change its approach to a minute error in the process by if-then-else coding.
Machine Learning
Machine learning is a wider term used for further categorized AI programs. Such as supervised learning, unsupervised learning and reinforcement learning. Machine learning goes from a simple pattern recognizing programs such as IBM’s Deep Blue which gathered much hype after beating Garry Kasparov in a one-on-one chess match in the 1990s, to the complex sentient machines who can learn human behavior and alter their interaction with human beings or other machines depending on the situation.
Machine learning goes beyond what today’s machines are capable of doing. To determine whether a machine has learned the pattern by decent coding or have the machine really evolved enough to generate its own answers and patterns, Alan M. Turing introduced a way he referred to as “imitation game”. Which is more commonly known as the Turing test? This test involves a computer, a human being, and an interrogator. The interrogator asks a bunch of questions from both the computer and the human and the computer has to convince the interrogator that it is a real human. This way of testing can help recognize a human and a machine as most of the programs such as Weizenbaum’s Eliza and Colby’s paranoid chatterbot Parry can have a conversation as a human being but once in a while they slip in out of context sentences.
Turing believed that machine learning holds enough potential that by the year 2000, an average interrogator would have less than 70% chance to distinguish between a man and a machine in a five-minute session. According to that estimation, we are still way behind.
Machine learning goes from limited memory programs like the RAM of your computer or self-driving cars which stores temporary data for situation analysis to a more advanced term of self-aware AI which is a machine that is aware of itself and its surroundings and interacts accordingly. A fully self-aware AI does not exist yet.
Some of the earlier categories which served as the prototypes of modern AI are:
o Natural language processing NLP
o Speech and audio processing
o Image and video processing
o Sensor and control processing
o Machine vision
o Robotics
Pros and Cons of AI Implementation
Artificial Intelligence has made human lives better and easier but it also comes with a couple of risks and extra safety measures. An idea of a being that is a hundred times more capable of doing whatever we as human beings can do, is dreamy and utopian for some and concerning and scary for others. If we indulge in benefits from AI implementations, we can also not ignore the risks that come with the increased integration of robots in our daily lives.
Pros
Increased quality in healthcare services is the biggest benefit of AI implementation. With faster and accurate data mining, an AI program like IBM Watson delivers error-free and fast answers. This sort of program can significantly improve healthcare services by helping both the doctors and the patients.
AI in business has increased the efficiency and accuracy of repetitive tasks that could easily be automated. Many companies are providing artificial intelligence as a service with AIaaS platforms.
Human safety has increased with the implementation of AI sensors.
Cons
Some of the biggest cons of AI technology is definitely fake audio and video generations. Many software is capable of providing fake conversations that never took place between two parties and even fabricated data required for the existence of a person like fingerprints and a face identity. “This person does not exist” is a famous website that generates real human faces which do not even exist, this gives the idea of how far into the “imitation game” we really are.
Top comments (0)