Popularly known through its acronym AI, Artificial Intelligence refers to the intelligence demonstrated by machines. Alternatively, it is also called Machine Intelligence or MI. The never-ending possibilities and innovations in AI have always been a source of fascination since its inception. The evolution of AI has been such that the significance of its presence today it is seemingly ‘absence’. In other words, it has weaved into the fabrics of society so well that it almost seems natural. Some functions performed by this remarkable technology today does not seem extraordinary because all of us have got used to it. Few examples of these are spam filters, online shopping recommendations, face recognition and the like.
The roles AI perform are varied and depends on the users’ requirements. Fundamentally there five ways in which AI can be used. The list below are the 5 A’s of Artificial Intelligence that stands for the 5 varied roles AI basically performs today.
‘Siri, book two tickets for a land at 3 today’ ‘Alexa, play Adele’s Skyfall’. Most of us are familiar with these dialogues and in fact quite often have conversations along these lines with our chatbots quite often. These are examples of Assistant AI’s which works by acting on the commands given to them. Assistant AI’s have integrated into our daily lives through our smartphones. Users can interact with this type of AI in their natural language and if any clarification is required, the former may ask questions in turn. The only drawback with AI as an assistant is the lack of predictive functionality, ie it cannot anticipate what might happen.
AI is sometimes used as an advisor or a consultant, doling out suggestions, tips or alternate ways of doing things. It differs from Assistant AI in the way it responds. When an advisor AI responds, it brings new insights into the dialogue rather than executing orders or merely asking for clarification like assistant AI. An enhanced assistant AI, it possesses prescriptive capabilities. Google Maps is an apt example of this type of AI; it gives the correct directions to a place, suggests the fastest route to that place (prescribes)and also predicts the estimated arrival time(predictive capabilities).
Together with AI, it is possible to do new things that cannot be otherwise accomplished solely by humans. With AI doing the monotonous and repetitive work which can be as complex or as simple it needs to be, humans can enjoy doing the best part of the fun aspects of the work. One important thing to remember is that AI performs the task in a semi-autonomous way; it does only its subtask and the human has to complete the rest of the task. An example can better illustrate this example. Let’s consider the example of translating from Hindi to English. An AI can translate an entire document from Hindi to English on its own. However, it still requires the final edit performed by a human. Similarly, AI used in law firms scans the hundreds of case law and prepares a draft document from it. This draft document can then be completed by an experienced lawyer.
AI as an authority is descriptive, predictive and prescriptive. People who work with Authority AI trusts the judgment and insights of the former. As AI has access to diverse and more volumes of data, it literally knows better than the humans who work with it. This fact gradually dawns on the human workers after working with authority AI in a matter of days. An example of this is the food apps which gives the delivery person clear instructions on when what and where the next delivery is required along with the evaluation of delivery performance and benefits.
The quintessential example of autonomous AI is self-driving cars. Self-driving cars only require initial instructions from the user, after which it can perform the entire function on its own. However, the users also have the option to interfere with the AI in any instance. Autonomous AI is thus an enhanced version of the autopilot system that is currently being used in trains and airplanes. Unlike the rigidly pre-programmed autopilots, autonomous AI is able to learn and improve its own performance.
There are a lot of speculations regarding two other predicted roles of AI ie Adversary AI & Affection AI. The former refers to AI emerging as a threat to humanity while the latter refers to building an effective relationship with AI. So far these two are limited to the pages of literature and imaginations. Hence it’s better to focus more on the possibilities of unethical use of AI by humans. With the bounds of progress that is affecting new Tech changes every day, there is a sure possibility that the 5 A’s of AI might be expanded to include much more. In the meantime, let’s first equip ourselves to deal with AI in the above mentioned five ways.