Understanding AI From Machine Learning to AI Agents
When I search the internet for an answer, I usually get a list of websites as search results. I have to visit multiple websites and spend time finding the information I need.
But that has changed dramatically. Today, AI can understand my question and provide the most relevant answer within seconds.
How does this happen? What powers these intelligent systems?
Key Takeaways
How AI impacts many of our daily activities.
We often use AI-powered technologies without even knowing it.
AI can improve productivity, automate repetitive tasks that typically require human intelligence.
AI can be categorized based on technology, capability, functional behavior, and business purpose.
What is AI?
AI is everywhere, from unlocking your mobile phone through facial recognition, powering Siri and Alexa, showing ads based on your searches, personalizing social media feeds and even enabling self driving cars and many more... But how do these systems know what you want? They learn from the data generated through your day-to-day activities and interactions.
AI stands for Artificial Intelligence. It is a technology designed to perform tasks that require human intelligence such as understanding language, recognizing patterns, making decisions and learning from experiences. AL learns from data to make decisions and improve its performance over time. The more relevant data it has, the better it can learn and become smarter.
Different Types of AI
AI can be categorize based on the following,
Technology
Capability
Functional Behavior
Business Purpose
Types of AI based on Technology
All includes many technologies that help machines learn, think and make decisions.
Machine Learning (ML)
ML is a Core branch of AI.
It allows machines to recognize patterns and make decisions or predictions based on data without programming each time, whereas traditional AI follows fixed rules.
Examples: Netflix content suggestions based on viewing history, Google Translate, Image Scan System, X-ray Analysis System etc.
Deep Learning (DL)
DL is a subset of Machine Learning that imitates the structure of the human brain using multi-layered artificial neural networks to understand and solve complex problems.
It is a key to modern AI applications.
Natural Language Processing (NLP)
Enables systems to understand, interpret, and respond to human language.
NLP needs both ML and DL.
Allows systems to process, understand text and speech to facilitate communication between humans and machines.
Computer Vision
Enables systems to interpret and analyze visual images and videos.
Supports object detection, facial and gesture recognition, vehicle tracking, smart surveillance etc by using DL techniques.
Types of AI based on Capability
AI is divided into different types based on how much human-like intelligence it can mimic and what kinds of tasks it can perform.
Artificial Narrow Intelligence (ANI)
Also known as Weak AI.
Designed to perform specific tasks or solve specific problems within a defined scope and it cannot perform tasks outside its specialized area.
ANI is widely used AI in our day to day activities.
Examples: Virtual Assistants such as Siri and Alexa, Recommendation systems used by streaming platforms like Netflix and Spotify, Facial Recognition systems, Chatbots, Speech -to - text transcription models, Fraud detection systems used by Banks and many more.
Artificial General Intelligence (AGI)
Also known as Strong AI.
Ability to think, learn, and apply knowledge across different tasks, just like humans.
It can learn from previous experiences and adapt to new challenges without requiring human intervention.
Currently AGI remains a theoretical and research concept. Researchers are trying to build machines that can handle different tasks on their own like humans.
Artificial Super Intelligence (ASI)
Capable of thinking, innovating, and reasoning at a level beyond what humans can achieve.
As with General AI, Super AI is still a concept and has not been created yet. Its development could bring many challenges and concerns about how it should be used and controlled.
Types of AI based on Functional Behavior
This classification is based on the way it processes and learns data, and responds.
Reactive Machines AI
The most basic form of artificial intelligence.
Do not have the ability to store data or learn from past experience.
Designed to react in real time, making them effective for straightforward tasks.
Examples: IBM Deep Blue - the chess playing computer that could analyze possible moves but lacked memory and learning capabilities, As well as Basic Spam Filters and Traffic Signal Control Systems that uses real time data, etc..
Limited Memory AI
Ability to use past data to improve its predictions or performance to some extent.
During development , machine learning models learn from data. After deployment, they usually use what they have learned and do not continue learning automatically unless they are specifically designed or retrained to do so.
Examples: Self-driving cars, Customer Service Chatbots, Smart Home Devices, Industrial Robotics and many more.
Theory of Mind AI
Future type of artificial intelligence that will be able to understand human emotions, thought, intentions, behavior and respond accordingly.
Unlike most current AI systems that work only based on commands and data, this type of AI aims to interact in a more human-like way.
Still in the research stage and has not yet been fully developed.
Examples: Current Virtual Assistants cannot truly recognize or understand human emotions and respond based on them.
Self-aware AI
Currently only a theoretical concept and has not been developed yet.
Ability to understand their own existence, make independent decisions and possibly have thoughts or awareness similar to humans.
Types of AI based on Business Purpose
This category is based on how AI is used in business use cases and workplace adoption.
Generative AI
Also Known as GenAI.
Generates Original content such as text, images, audio, code, videos, presentations and more based on patterns learned from large amounts of training data.
Creates new content based on user prompts and questions.
Examples: Drafting emails, debug code, generating newsletters etc.
Predictive AI
Using statistical analysis and machine learning to identify patterns, anticipate behaviors and forecast upcoming events.
Organizations use predictive AI to predict potential future outcomes, causation, identify risks, support decision making and many more.
Examples: Weather forecasting , Traffic prediction and Customer Behavior analysis etc.
Assistive AI
Designed to help people work faster and smarter by offering suggestions, automating repetitive tasks, and providing on demand support within the applications and tools.
Examples: Navigation assistance, Grammar corrections, Email auto-completion, Summarization and many more.
Conversational AI
Able to simulate human conversation by using natural language processing (NLP) and machine language(ML).
Examples: Virtual Agents, Customer Service applications, Mobile Devices and Smart Speakers and many more.
Agentic or Autonomous AI
Agentic AI uses an LLM as its reasoning engine to independently plan, use external tools, and execute multi-step workflows to achieve a specific goal.
Requires minimal human intervention.
It is one of the fastest-growing areas of AI.
Examples: Automated fraud detection, software code generation and debugging, data analysis for supply chain management, autonomous task execution and many more.
AI is transforming the way we live and work. From simple chatbots to intelligent AI agents, its capabilities continue to grow every day. Understanding AI today will help us make the most of the opportunities it brings tomorrow.
