Introduction to AI and Chatbots

Introduction to AI

• Artificial Intelligence is the science and engineering dedicated to creating intelligent machines and computer programs

AI Purposes

• AI serves to enhance human intellect through computer capabilities, similar to how motors amplify physical strength
• Robotics and expert systems are significant components of this enhancement
• AI also functions to enable computers to comprehend human thought processes

Chatterbots

• A chatterbot, chatbot, or chat bot simulates intelligent conversations via both auditory and textual channels
• The main goal is simulation is to emulate human-like conversation to a degree a user might be tricked into thinking they are talking to a human
• Can be classified by: intelligence, purpose, and technology

Chatterbot Intelligence Classifications

• Rule-Based Chatterbots: Operate using predefined rules, decision trees, or keyword recognition and are limited in complex scenarios
• AI-Powered (Machine Learning) Chatterbots: Incorporate Natural Language Processing (NLP) and Machine Learning (ML) to interpret intent and context and improve with data and interactions
• Hybrid Chatterbots: Combine rule-based with AI learning for a better user experience to handle structured and open-ended conversations

Examples of Specific Chatterbots

• ChatGPT (OpenAI): Can understand and generate human-like responses
• Google Bard (Gemini AI): Google’s AI chatbot is designed for creative writing and information retrieval
• Claude (Anthropic): An AI assistant optimized for ethical and safe interactions
• IBM Watson Assistant: An example of an AI used in customer service and business analytics
• Mitsuku (Pandorabots): A chatbot designed to engage in human-like conversations

Chatterbot Purpose Classifications

• Customer Support Chatterbots: Handle many FAQs, complaints, and general inquiries
• Examples: Erica (Bank of America) helps users manage their finances and Eno (Capital One) monitors transactions and alerts users to suspicious activity
• Personal Assistant Chatterbots: Help users with scheduling, reminders, and general queries
• Examples: Siri (Apple) is an AI-powered voice assistant and Alexa (Amazon) is able to provide smart home control and assistance
• Google Assistant is also integrated into Android devices
• Social & Companion Chatterbots: focus on casual conversations, mental well-being, and companionship
• Examples: Replika provides emotional support and friendship and Woebot is a mental health chatbot
• Healthcare Chatterbots: Provide medical guidance, symptom analysis, and mental health support
• Examples: Ada Health is an AI-powered symptom checker and Babylon Health assists with medical consultations
• Education & Learning Chatterbots: Help students with learning, tutoring, and language practice
• Examples: Duolingo AI Chatbot is a language-learning assistant and Socratic (Google) is an AI tutor for solving homework problems
• E-commerce & Shopping Chatterbots: Assist with product recommendations, purchases, and order tracking
• Examples: H&M Chatbot and Sephora Virtual Artist are fashion and beauty focused

Chatterbot Interaction Method Classifications

• Text-Based Chatterbots communicate through text in messaging apps, websites, or apps
• Examples: Facebook Messenger Bots helps businesses with customer service and Slack Bots help with team collaboration
• Voice-Based Chatterbots are designed to recognize and respond to voice commands
• Examples: Cortana (Microsoft) a voice AI assistant, and Google Nest Hub
• Multimodal Chatterbots can handle both text and voice, and sometimes images and videos
• Examples: Gemini AI (Google Bard) accepts voice and image inputs, and Meta AI is integrated into Meta apps such as Facebook and Instagram

History of AI: Ancient Era (Before 1900)

• ~300 BCE: Greek mythology has Talos, a giant mechanical man and ancient automata are built in China, Egypt, and Greece
• 1495: Leonardo da Vinci designs a mechanical knight that can move
• 1642: Blaise Pascal invents the Pascaline, a mechanical calculator
• 1854: George Boole introduces Boolean algebra, which lays the groundwork for binary systems

History of AI: Early Computational Foundations (1900–1950)

• 1936: Alan Turing introduces the Turing Machine concept, which is a universal computing device
• 1943: Warren McCulloch and Walter Pitts create a model of artificial neurons
• 1950: Alan Turing introduces the Turing Test

History of AI: Birth of AI (1956–1970)

• 1956: The term “Artificial Intelligence” is coined at the Dartmouth Conference by John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon
• 1958: John McCarthy develops LISP
• 1966: ELIZA, an early chatbot by Joseph Weizenbaum, simulates a psychotherapist
• 1969: Shakey the Robot, the first general-purpose mobile robot, is created at Stanford

History of AI: First AI Winter (1970s)

• 1970s: AI progresses slowly due to the high cost and limited power of early computers
• 1972: Prolog, a logic-based AI programming language, is developed

History of AI: Expert Systems Era (1980s)

• 1980: AI research improves with Expert Systems
• 1987: AI funding declines again as expert systems prove expensive and limited

History of AI: Machine Learning and Early Modern AI (1990s–2000s)

• 1997: IBM’s Deep Blue defeats Garry Kasparov
• 1999: AI enables recommendation systems and personalized content
• 2000s: Machine learning gains momentum boosted by vast amounts of data and improved hardware

History of AI: Deep Learning Revolution (2010–2019)

• 2012: AlexNet wins the ImageNet competition, popularizing deep learning and CNNs
• 2016: Google DeepMind’s AlphaGo defeats Lee Sedol
• 2018: OpenAI's GPT advances natural language processing

AI in Everyday Life (2020–Present)

• 2020: AI is found in virtual assistants, self-driving cars, and medical diagnosis
• 2021: OpenAI’s GPT-3 is created
• 2022: DALL·E and other generative AI models create images from text descriptions
• 2023–2025: Focus on continued advancements in AI ethics, AGI research, and human-AI collaboration

AI Agents and Rationality

• Intelligent Agent: A program that uses its environment, user input, and experiences to act and make decisions and can be used to perform tasks autonomously or in response to user prompts
• Simple reflex agents: Operate on "if-then" principles for simple tasks in predictable environments
• Goal-based agents: Agents who use goals to describe desirable situations
• Utility-based agents: Evaluate actions using possible outcomes for the most beneficial choice
• Learning agents: Adapt and learn automatically from past experiences
• Agents: Perceive the environment through sensors and act upon it through actuators
• Rationality: Performance measuring success, agents' prior knowledge, and percept sequence to date
• Rational Agent: Should chose an action expected to maximize its performance
• Rational Choice Theory: States that individuals use rational calculations to make choices that aligned with their objectives
• Concept of Rationality: guides people to make good choices or decisions from an economics point of view
• Autonomy: Ability to make decisions and act independently and in accordance with beliefs and values.
• No autonomy: Ignores environment/data
• Complete autonomy: Acts randomly/no program
• Ideal: Design agents to maximize autonomy

Key Elements for Intelligent Agent Design

• Performance measure
• Environment
• Actuators
• Sensors

AI Environments & Uncertainty

• Environment types include Fully observable (vs. partially observable), Deterministic (vs. stochastic), Episodic (vs. sequential), Static (vs. dynamic), Discrete (vs. continuous), Single agent (vs. multiagent)

Environment Observability

• Fully observable: The agent knows all information about the environment
• Partially observable: The agent has partial information and based on incomplete knowledge infers hidden aspects

Environment Predictability

• Deterministic: the environment is certain and predictable
• Stochastic: The environment is random and uncertain

Environment Continuity

• Episodic: events happen at irregular intervals
• Sequential: events happen in a specific order
• In AI, episodic environments are independent, while sequential environments are dependent

Environment Stability

• Static: The environment is fixed or stationary
• Dynamic: The environment is energetic and forceful

Environment State Number

• Discrete: Has a finite number of possible states and actions
• Continuous: Has an infinite number of possible states and actions

Number of Interacting Agents

• Single agent: Operates independently
• Multi-agent: Multiple agents interact and collaborate

Uncertainty in AI

• AI systems face uncertainty due to incomplete or ambiguous information
• Types of AI uncertainty include:
  • State uncertainty: Lack of information to predict market trends and changes
  • Effect uncertainty: Difficulty predicting AI's impact on business

Examples of AI Reasoning Under Uncertainty

• Medical Diagnosis: AI systems assist in diagnosing diseases
• Autonomous Vehicles: AI must make quick decisions in uncertain environments

Reasons to Study AI

• AI revolutionizes technology, enabling AI and machine learning to enhance business technologies
• AI boosts career prospects by equipping individuals with expertise sought by tech companies
• AI knowledge facilitates adaptation across various industries, ensuring relevance in healthcare, finance, and more
• AI studies provide versatile skills leading to opportunities and high earning potential, averaging between 100,000to100,000 to 100,000to150,000 for mid-level professionals in the US
• AI is a key skill for the 21st century, automating taskutubes and creating new job roles
• AI allows for connecting with global experts in online forums, group projects, and seminars
• AI knowledge is relevant to various interests, from developing AI tools to creating sustainable solutions
• AI fosters innovation and creative problem-solving beyond traditional data analysis
• AI enhances decision-making skills through data insights, benefitting business strategies and customer relations
• AI adoption supports critical environmental issues through AI technologies and sustainable practices
• AI is crucial for creating customized experiences in marketing and e-commerce.
• AI bolsters cybersecurity by developing advanced security systems using machine learning
• AI-driven research advances medical research and enhances space exploration using AI-powered robots
• AI literacy prepares everyone for a future where AI is integrated into daily life
• AI education encourages ethical solutions that are fair
• AI enhances knowledge across computer science, mathematics, psychology, and linguistics
• AI is shaping smart cities and infrastructure, by intelligently managing resources and urban environments
• AI revolutionizes diagnosis and patient care, improving healthcare outcomes
• AI enhances value in the gig economy, providing freelance opportunities in strategies and bespoke AI solutions
• AI can transform education through personalized tutoring and educational tools
• AI is pivotal in optimizing operations across industries by automating routine taskutubes and complex data entries, crucial for banking and financial sectors
• AI impacts creative media, blending traditional and new content. AI tools assist music industry professionals to generate new content and visual artists in digital art, personalizing media experiences
• AI will improve disaster response with better predictive capabilities for events like hurricanes and floods
• AI tools allow in financial markets the speedy analysis of fast data sets that humans can't compete with
• AI promotes public sector analytics by automating administration taskutubes to improve public awareness
• AI contributes to agricultural processes through analytical and data reading that informs procedures on waste reduction
• AI has developed real time translation systems that breaks down communication barriers through verbal translation
• AI promotes streamlined retail customer analytics to personalize interactions within the retail sector
• AI reshapes the entertainment industry, enabling more interactive visual effects to decrease costs. Algorithmic analysis adjusts decisions on gaming environments that adapts to players
• AI supports fairness transparency, creating equity against societal impact of technological bias