AI Notes

Foundational Concept of AI

  • Intelligence:
    • Ability to interact with the real world by making inferences, understanding, and reacting.
    • Ability to deal with uncertainties through proper planning and decision making.
    • Ability to learn and adapt continuously, like a baby.

What is Artificial Intelligence?

  • Artificial: Man-made, not occurring naturally.
  • Intelligence: (As defined previously).
  • Artificial Intelligence (AI): When a machine mimics human traits, such as decision-making, predicting the future, learning, and self-improvement.

Applications of AI

  • Voice-based assistants (e.g., Google Assistant, Alexa, Siri).
  • Google Maps.
  • Recommendation systems (e.g., Amazon, YouTube, Netflix).
  • Chatbots.
  • Snapchat filters.

Domains of AI

Statistical Data

  • AI domain related to data systems and processes.
  • The system collects numerous data, maintains datasets, and derives meaning from them.
  • Information extracted can be used for decision-making.
  • Examples: Price comparison websites, music recommendations.

Computer Vision

  • AI domain depicting a machine's capability to acquire and analyze visual information, then predict decisions.
  • Process: Image acquisition, screening, analyzing, identifying, and extracting information.
  • Input: Photographs, videos, pictures from thermal or infrared sensors, etc., converted into computer-readable language.
  • Objective: Teach machines to collect information from pixels.
  • Examples:
    • Agricultural Monitoring.
    • Surveillance Systems: Monitoring public spaces, detecting suspicious activities, tracking individuals/vehicles, providing real-time alerts.

Natural Language Processing (NLP)

  • Branch of AI dealing with interaction between computers and humans using natural language.
  • NLP attempts to extract information from spoken and written words using algorithms.
  • Objective: To read, decipher, understand, and make sense of human languages.
  • Examples:
    • Email filters: Identifying spam messages.
    • Machine Translation: Systems like Google Translate and Microsoft Translator analyze sentence structure and semantics to translate text between languages automatically.

Ethical Frameworks for AI

  • Frameworks provide a structured approach to problem-solving, ensuring all relevant factors are considered.
  • Help ensure choices do not cause unintended harm.
  • Provide a systematic approach to navigating complex moral dilemmas.
  • Help make well-informed decisions aligning with values and promoting positive outcomes for stakeholders.

Why Ethical Frameworks for AI?

  • To prevent bias and unwanted outcomes in AI solutions (e.g., biased hiring algorithms).
  • Ensure AI makes morally acceptable choices.
  • Avoid unintended outcomes during AI solution development.

Types of Ethical Frameworks

  • Sector-based.
  • Value-based.
    • Rights: Human life valued over others.
    • Utility: Offers more good than harm.
    • Virtue: Actions consistent with beliefs.

Sector-based Frameworks

  • Tailored to specific sectors or industries.
  • Bioethics: Focuses on ethical considerations in healthcare (patient privacy, data security, ethical use of AI in medical decisions).
  • Applicable to finance, education, transportation, agriculture, governance, and law enforcement.

Value-based Frameworks

  • Focus on fundamental ethical principles and values guiding decision-making.
  • Reflect moral philosophies.
  • Assess the moral worth of actions.
  • Types:
    • Rights-based: Prioritizes human rights and dignity, respecting individual autonomy and freedoms (ensuring AI systems do not violate human rights).
    • Utility-based: Maximizes utility or overall good, aiming for the greatest benefit while minimizing harm.

Virtue-based Frameworks

  • Focuses on the character and intentions of individuals involved in decision-making.
  • Asks if actions align with virtuous principles (honesty, compassion, integrity).
  • In AI: Consider whether developers, users, and regulators uphold ethical values throughout the AI lifecycle.

Bioethics

  • Ethical framework used in healthcare and life sciences.
  • Deals with ethical issues related to health, medicine, and biological sciences.
  • Principles:
    • Respect for Autonomy.
    • Do no harm (Non-maleficence).
    • Ensure maximum benefit for all (Beneficence).
    • Give justice.

Four Principles of Bioethics

  • Respect for autonomy: Enabling users to be fully aware of decision-making (users should know how an AI algorithm functions).
  • Do no harm (Non-maleficence): Avoiding harm to anyone (human or non-human); choose the path of least harm.
  • Maximum benefit (Beneficence): Focus on providing the maximum benefit possible.
  • Justice: Distribute benefits and burdens fairly across people, irrespective of background.

Ethical Principles

  • Non-maleficence: Avoiding causing harm or negative consequences; minimizing harm.
  • Maleficence: Intentionally causing harm or wrongdoing.
  • Beneficence: Promoting and maximizing well-being and welfare; taking actions that produce positive outcomes.

AI Project Cycle - 6 stages

  • Problem Scoping
  • Data Acquisition
  • Data Exploration
  • Modeling
  • Evaluation
  • Deployment

Test Yourself (Answers from Transcript)

  1. B) To understand the aim and objective of the project
  2. A) Two domains
  3. B) Analyzing data to extract insights
  4. D) Converting digital visual data into computer-readable language
  5. C) Dealing with the interaction between computers and humans using natural language
  6. B) Step-by-step guidance
  7. B) Into sector-based and value-based frameworks
  8. C) Aligning actions with ethical principles and beliefs
  9. A) Prioritizing human rights and dignity, valuing human life over other considerations
  10. B) Healthcare and life sciences
  11. A) Both Assertion and Reasoning are true, and Reasoning is the correct explanation of the Assertion.
  12. A) Both Assertion and Reasoning are true, and Reasoning is the correct explanation of the Assertion.

Reflection Time (Topics from Transcript)

  1. Outline the main steps in the AI Project Cycle briefly.
  2. What roles does computer vision play in agricultural monitoring systems?
  3. Mention the factors which knowingly or unknowingly influence our decision-making.
  4. What is the necessity for Ethical Frameworks in AI development?
  5. Mention the key characteristics of sector-based frameworks.
  6. What do you mean by Bioethics?
  7. What is Natural Language Processing? Explain any two real-life applications of NLP.
  8. How do value-based frameworks contribute to ethical decision-making by emphasizing fundamental principles and values?

Various Abilities Involved in Intelligence

  • Musical Intelligence
  • Intrapersonal Intelligence
  • Interpersonal Intelligence
  • Naturalist Intelligence
  • Mathematical Logic Intelligence.
  • Linguistic Intelligence
  • Spatial Visual Intelligence
  • Kinesthetic Intelligence
  • Existential Intelligence

Intelligences

  • Mathematical Logical Reasoning:
    • Ability to understand and use numerical symbols, abstraction, and logic.
  • Linguistic Intelligence:
    • Language processing skills (understanding, writing, verbally).
  • Spatial Visual Intelligence:
    • Ability to perceive the visual world and relationships between objects.
  • Kinesthetic Intelligence:
    • Ability to use one's limbs in a skilled manner.
  • Musical Intelligence:
    • Ability to recognize and create sounds, rhythms, and sound patterns.
  • Intrapersonal Intelligence:
    • Level of self-awareness (strengths, weaknesses, feelings).
  • Existential Intelligence:
    • Relating to religious and spiritual awareness.
  • Naturalist Intelligence:
    • Ability to process information on the environment.
  • Interpersonal intelligence
    • Ability to communicate with others by understanding other people's feelings & influence of the person.

Key Points

  • Any machine trained with data that can make decisions/predictions is considered AI.
  • 'Training’ is crucial; not all “smart” devices are AI-enabled.

What is Not AI?

  1. Automatic Washing Machine:
    • Requires human input to select parameters; example of automation.
  2. Air Conditioner controlled via IoT:
    • Needs human interaction; example of Internet of Things (IoT).
  3. Automation via Sensors:
    • If the machine isn't trained with data, it's not AI.

What is AI and Why?

  • Google Search Engine:
    • AI turns it into an intelligent search engine with direct answers, voice/image searches, and deep learning.
  • Voice Assistant:
    • Uses AI to recognize spoken words.
    • NLP capabilities (speech-to-text).
    • Learns via ML algorithms, improving performance over time.

Gaming World:

  • Games use Machine Learning (ML) algorithms to understand human patterns and adjust difficulty levels.
  • Adaptive AI adjusts game difficulty based on player skill.

Importance of Data in AI Devices

  • Machines should be trained with accurate data.
  • More accurate data leads to better predictions.

AI and IoT Blended

  • Internet of Things (IoT): Network of physical objects embedded with sensors + software for data exchange.
    • Goal: self-reporting devices in real-time, improving efficiency.
    • Ex: Smartwatches, Fitbit, smart home gadgets

AI and IoT

  • AI-enabled IoT creates intelligent machines that simulate smart behavior and support decision-making with minimal human interference.
  • IoT deals with devices interacting via the internet; AI enables devices to learn from data and experience.