The Non-Technical Guide to Artificial Intelligence
- Jun 4, 2021
- 6 min read
Updated: Jun 7, 2021
Are you struggling to grasp the idea behind Artificial Intelligence? Or trying to search for the layman explanation of Artificial Intelligence? Well, if you are, then this is your one-stop destination for AI made easy.
Artificial Intelligence (AI) is a broad term used in multiple contexts, often making no sense at all. There are a ton of different domains and sub-topics related to Artificial Intelligence; from computer algorithm in machine learning (ML), manipulation and interpretation of languages in Natural Language Processing (NLP) to an endless list of enigmatic terminologies and fields in AI. In short, the terms that go over our non-technical-heads.
There is no single definition for AI, but a plethora of different, vague interpretations. The easiest definition that I would coin is “Ability of a computer program to ‘think’ and ‘learn’, similar to that of humans”. To cut that down, AI is simply a simulation of human intelligence in machines.
To simplify the document, here are some important "non-technical" definitions:
Artificial Intelligence (AI): Ability of a computer program to 'think' and 'learn', similar to human capabilities. AI is simply a simulation of human intelligence in machines.
Machine Learning (ML): Type of AI, that enables self-learning methodology from data to apply the learning without the need for human intervention. With ML, computers self-learn without being explicitly programmed, i.e., it doesn't require teaching. Machines receive data and use it to learn.
Natural Language Processing (NLP): Branch of AI that enables computers to understand, comprehend, interpret, and manipulate the human language. In short, NLP helps machines "read" by simulating the human ability to understand language, as per expert.ai.
Augmented Reality (AR): Augmented Reality and Artificial Intelligence are closely related to each other. AR technology enables the interaction between digital and physical environments, i.e., it allows physical environments to be simulated digitally.
Artificial Intelligence and its aspects surround us throughout our lives, and we don’t even realize it. We all use Google’s search engine to find things online. Have you ever been shocked to see that what you want to search pops up in the options in the search bar? Worry not, it is just Google’s search engine AI in action. It simply understands your queries from the few words, and your history, and automatically generates a response to the same, helping you find the nearest Starbucks quickly.
E-commerce and media streaming
AI is used in eCommerce and entertainment websites for recommendations. Most commonly, it is seen in Amazon’s and Netflix’s recommendation system to match your products or shows and movies. Such systems use Machine Learning (ML). ML enables self-learning methodology from data and applies the data in real-life concepts without having human intervention. In Netflix for instance, the ML model recognizes the kinds of shows users watch and accordingly suggests similar show categories. This creates a loop – show/movie watched by users, similar category shows seen by millions on Netflix, recommended to users.
Travel and Navigation
AI is used in Google Maps. Just think, don't maps rely on GPS? The concept of Live-view in Google maps is dependent on AI, scanning billions of street and local images, to help understand the user’s orientation and placement, even inside a mall or an airport. Say goodbye to getting stuck at airports! Google maps also uses ML models and Augmented Reality technology to do the same. As more and more people use Google maps to travel around the world, ML models take information and use it to enhance the navigation experience to predict traffic around a route, and even suggest an estimated time of arrival. Augmented Reality is a technology closely related to AI, which lets users superimpose images, texts, and sound over a real-world simulated environment.
Autocorrect and Language
Believe it or not, even autocorrect and software like 'Grammarly' use AI models. "Grammarly" uses a form of AI system called Natural Language Processing (NLP). For most of the human language processing tasks, NLP is used. In a nutshell, NLP helps computers comprehend, understand and analyze and manipulate the human language – which is the main task of Grammarly and autocorrect on smartphones, 'Microsoft word' and others.
Social Media
Now to the favorite example - Social Media! Be it Instagram, Facebook, or Twitter, every social networking service uses some kind of AI to help its users. Let's have a look at how AI is integrated with social media:
Instagram and Facebook use AI models to generate feeds based on what users prefer, posts they like, people they follow, and food tagging. Using these criteria, the AI finds the most suitable content to be placed on top of the feed which ensures you spend more time online. Both also use AI to provide more user engagement.
Twitter utilizes AI to identify what tweet recommendations should be suggested on user timelines and aims to highlight the most relevant tweets for every user. It also uses AI to identify cyberbullying, cyber-aggression, and toxic behavior. This is done using NLP models.
LinkedIn leverage AI to predict suitable candidates for certain job positions. It also uses AI to recommend the ones who are looking for jobs online.
Aspects of Artificial Intelligence
Artificial Narrow Intelligence (ANI)
Have you ever told Siri to set an alarm to check the Pizza in the oven? Or told Amazon’s Alexa to put on Netflix? These tasks are accomplished by Artificial Narrow Intelligence (ANI). This kind is the current form of AI, which can execute specific focused tasks. ANI uses a single subset of cognitive abilities to perform such tasks. This kind of AI has surpassed humans in multiple simple fields such as gaming and can play mindful games like chess, beating world champions. These AI systems play thousands and thousands of games and store every single move played to continuously learn and become a pro.
Even advanced AI such as handwriting recognition and Self-Driving cars are examples of ANI
Artificial General Intelligence (AGI)
Artificial General Intelligence (AGI), is a technology that is at par with human intelligence or the human mind. AGI allows machines and algorithms to carry out multiple tasks with ease and efficiency. A very common example would be Jarvis from Marvel’s Iron Man. Jarvis is seen to easily comprehend language, understand, think, and behave similarly to a human being. AGI brings the advantage of flexible thinking and reasoning with the computational advantages of being a machine. Technically, theories suggest that anything a human can do, AGI models can as well. Nevertheless, for AI to be human, it must:
Transfer learning from topic to topic – infused knowledge between fields
Gain consciousness
Have common sense
Be able to work and collaborate with others
These things are, however, very difficult for AGI models to conceive at earlier stages. AGI systems are designed with the human brain as a reference and duplicating the super-complex structure is only theoretically possible at the moment. Thereby, only speculative whether AGI will be possible in the future or not.
Artificial Superintelligence (ASI)
Lastly, the highest level that Artificial Intelligence can reach – Artificial Superintelligence or ASI. Exceeding all levels of human intelligence, in all possible areas, be it creativity, logic, knowledge, and problem-solving – ASI has gotten it covered. It surpasses human intelligence, making them the smartest species on Earth. At the moment, humans are the primary species, solely because of our intelligence and thereby our capabilities of performing “intelligent tasks”. The concept of ASI is still hypothetical and debates happen across the globe whether it could be accomplished, and by when? ASI would have the capabilities to find answers to most of our unanswered questions – perhaps even find a cure to cancer, dementia, and many more.
AGI and ASI systems are designed with the human brain as a reference and duplicating the super-complex structure is only theoretically possible at the moment. But, even though the creation of such systems is speculative and debatable, with the rapid rate of newer AI capabilities, we might be close.
Where have we reached in the scope of Artificial Intelligence?
Technologies are rapidly advancing. We started from replacing horse and human power with steam engines in the first industrial revolution to now Artificial Intelligence, and robotics in fourth, and soon to be the fifth industrial revolution. It took us almost 1 century to switch from steam-based machines to electrical mass-production, while it only took a few decades to get into advanced computing and AI. We are constantly and rapidly moving forward. The younger generations have seen a sudden change in the way we live, and it is yet to change further. If we continue this pace, we will be able to reach AGI by 2050, as per experts like Ray Kurzweil. However, there are speculations about the same - while some say it is not possible, others say it will be there within a few decades.
Artificial Intelligence is expanding globally in a rapid manner. Even though we might be decades away from AGI and ASI systems, we are slowly adopting the simpler narrow AI systems in our lives and soon, we will be ready to adopt AI in all possible contexts. AI is already entering industries and becoming a general-purpose technology. Understanding the concept of AI and ML is deeply important as it holds the future of not only tech but all of us. AI holds promise and shines light on newer human capabilities in the future.
Comments