0% found this document useful (0 votes)
4 views10 pages

Module 3 Notes

The document discusses design metaphors in AI, emphasizing the relationship between humans and machines, and the importance of human control in AI systems. It outlines two main goals of AI research: the science goal, which aims to create autonomous machines that replicate human intelligence, and the innovation goal, which focuses on enhancing human capabilities through AI assistance. The document also presents various design metaphors, such as intelligent agents, super tools, teammates, and tele-bots, highlighting their roles in AI development and deployment.

Uploaded by

DARK'S GAMING
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
4 views10 pages

Module 3 Notes

The document discusses design metaphors in AI, emphasizing the relationship between humans and machines, and the importance of human control in AI systems. It outlines two main goals of AI research: the science goal, which aims to create autonomous machines that replicate human intelligence, and the innovation goal, which focuses on enhancing human capabilities through AI assistance. The document also presents various design metaphors, such as intelligent agents, super tools, teammates, and tele-bots, highlighting their roles in AI development and deployment.

Uploaded by

DARK'S GAMING
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Human-Centered AI (BAI613)

MODULE 3
DESIGN METAPHORS IN AI
Topics: Introduction to Design Metaphors in AI, Science and Innovation Goals, Intelligent Agents,
Super tools, Teammates, Tele-bots, Social Robots, Active Appliances

1. Introduction to Design Metaphors in AI


Artificial Intelligence systems can be designed in different ways depending on how researchers
view the relationship between humans and machines. To explain and guide these design
approaches, researchers use design metaphors.

A design metaphor is a conceptual comparison that helps us understand how AI systems should
behave and interact with humans. It answers an important question:
Should AI replace humans, or should AI assist humans?

This question forms the foundation of modern AI design.

The “Bicycle for the Mind” Metaphor

Steve Jobs famously described computers as:

“A bicycle for our minds”

This metaphor means that computers should amplify human abilities, just like a bicycle amplifies
physical movement. The bicycle does not replace humans; instead, it helps humans move faster
and more efficiently while still remaining in control.

In contrast, some researchers envision computers as autonomous systems that operate like a
chauffeur-driven limousine, where humans simply provide the destination and the machine handles
everything automatically.

This raises an important design concern:

If machines operate autonomously, how do humans control them, supervise them, or correct them
when errors occur?

Aviation Example: Importance of Human Control

Modern aircraft provide an excellent example of balanced AI design.

Aircraft use automated systems such as autopilot, which can control speed, altitude, and
navigation. However, human pilots remain essential because unexpected situations may arise, such
as:

 Engine failure
HARSH VERMA 1
Human-Centered AI (BAI613)

 Weather emergencies

 Bird strikes

 Mechanical faults

In such situations, human intervention becomes critical.


This example demonstrates that fully autonomous systems may not always be reliable, and human
supervision remains necessary.

Role of Design Metaphors in AI Development


Design metaphors influence how AI systems are built and deployed in real-world environments.
They help determine:
 Level of machine autonomy

 Level of human control

 Safety and reliability

 User trust and acceptance

Design metaphors are especially important in critical domains such as:

 Healthcare

 Transportation
 Military systems

 Financial systems

 Industrial automation

The goal is to create systems that are both intelligent and safe.

2. Science and Innovation Goals of AI

AI research is driven by two major goals:

1. Science Goal

2. Innovation Goal
These goals represent two different philosophies of AI development.

2.1 Science Goal of AI

The science goal focuses on understanding intelligence itself and creating machines that can think
and act independently like humans.

HARSH VERMA 2
Human-Centered AI (BAI613)

Researchers pursuing the science goal aim to understand human intelligence and replicate it in
machines.

This involves studying human:

 Perception

 Reasoning

 Learning

 Decision making
 Emotional response

 Motor skills

The ultimate aim is to create intelligent machines capable of performing tasks autonomously.

Understanding Computational Agents

Science goal researchers focus on creating computational agents that act intelligently.

A computational agent is a system that can:

 Perceive its environment through sensors


 Process information

 Make decisions

 Take actions independently

These agents behave similarly to intelligent humans.

For example, AI systems such as AlphaGo learned to play complex games by analyzing patterns
and improving their performance through experience.

Examples of Science Goal Applications

Autonomous Vehicles

Self-driving cars use sensors, cameras, and AI algorithms to navigate roads without human
intervention.

They detect obstacles, recognize traffic signals, and make driving decisions independently.

The goal is to achieve complete autonomy.


Medical Diagnosis Systems

AI systems analyze medical images such as X-rays and MRI scans to detect diseases like cancer.

HARSH VERMA 3
Human-Centered AI (BAI613)

These systems attempt to match or exceed human diagnostic ability.

Speech and Language Understanding

AI systems such as virtual assistants can recognize speech, understand language, and generate
responses.

This demonstrates machine capability to mimic human communication.

Robotics

Autonomous robots can navigate environments, perform tasks, and make decisions without human
control.

Examples include warehouse robots and exploration robots.


Long-Term Goal: Artificial General Intelligence (AGI)

The ultimate science goal is to create Artificial General Intelligence (AGI).

AGI refers to machines that can perform any intellectual task that humans can perform.

Such machines would possess human-level intelligence and adaptability.

However, AGI remains a future research goal.

Characteristics of Science Goal Systems

Science goal systems are characterized by:


 Autonomy

 Independent decision making

 Self-learning capability

 Goal-directed behavior

These systems aim to function independently of human control.

2.2 Innovation Goal of AI

The innovation goal focuses on developing AI systems that assist humans rather than replace them.

This approach emphasizes Human-Centered AI (HCAI).


Instead of creating independent machines, innovation goal researchers focus on creating tools that
enhance human abilities.
Purpose of Innovation Goal

The innovation goal aims to:

HARSH VERMA 4
Human-Centered AI (BAI613)

 Improve human productivity

 Assist in decision making

 Automate repetitive tasks

 Enhance human performance


In this approach, humans remain in control, and AI functions as a support tool.

Examples of Innovation Goal Systems

Navigation Systems

GPS navigation systems provide route suggestions but allow humans to choose their path.

The AI assists but does not replace human decision making.

Medical Assistance Systems

AI systems help doctors analyze medical data, but doctors make final decisions.

This improves diagnostic accuracy while maintaining human control.

Writing Assistance Tools

HARSH VERMA 5
Human-Centered AI (BAI613)

Applications such as grammar checkers suggest corrections, but users decide whether to accept
them.

Virtual Assistants

Virtual assistants such as Siri and Alexa help users perform tasks like searching information and
controlling devices.

These assistants function as tools rather than autonomous decision makers.

Characteristics of Innovation Goal Systems

Innovation goal systems emphasize:

 Human control
 Assistance rather than replacement

 Enhanced productivity

 Safety and reliability

These systems empower humans rather than replacing them.

3. Combined Approach: Integrating Science and Innovation Goals

Modern AI design combines both science and innovation goals.

Machines handle tasks that can be automated reliably, while humans retain control over critical
decisions.

This combined approach provides both efficiency and safety.

Example: Digital Camera


Modern digital cameras automatically adjust focus and lighting using AI.

However, humans still control important aspects such as framing and timing.

This combination improves performance while preserving human creativity.

Example: Self-Driving Cars

Current self-driving systems assist drivers by:

 Maintaining lane position

 Detecting obstacles
 Preventing collisions

However, human drivers remain responsible for overall control.

HARSH VERMA 6
Human-Centered AI (BAI613)

Example: Industrial Control Systems

Industrial systems use AI to monitor and optimize performance, but human operators supervise
operations.

This ensures safety and reliability.

4. Intelligent Agents and Super tools

This is the first pair of design metaphors.

Intelligent Agents
Intelligent agents are autonomous systems capable of independent operation.

They can perceive environments, make decisions, and act without human control.

Examples include:

 Autonomous robots

 Game-playing AI systems

 Autonomous vehicles

These systems attempt to replicate human intelligence.


Super tools

Super tools are AI systems designed to enhance human capabilities.

They function as powerful tools controlled by humans.

Examples include:

 Search engines

 Data analysis software

HARSH VERMA 7
Human-Centered AI (BAI613)

 Writing assistance tools

These tools amplify human abilities without replacing humans.

5. Teammates and Tele-bots


This is the second pair of design metaphors.

Teammates

In this metaphor, AI systems act like human teammates.

They collaborate with humans and assist in tasks.

However, machines lack true human understanding and responsibility.

This metaphor can sometimes create unrealistic expectations.

Tele-bots

HARSH VERMA 8
Human-Centered AI (BAI613)

Tele-bots are remotely controlled machines operated by humans.

Examples include:

 Surgical robots

 Military drones
 Space exploration robots

These systems extend human capabilities while maintaining human control.

6. Assured Autonomy and Control Centers

This is the third pair of design metaphors.

Assured Autonomy

Assured autonomy refers to fully autonomous systems capable of independent operation.

Examples include autonomous vehicles and automated trading systems.

However, fully autonomous systems can fail unpredictably.


Control Centers

Control centers allow humans to monitor and supervise automated systems.

Examples include:

 Air traffic control

 Mission control centers

 Industrial monitoring systems

These systems ensure safety through human oversight.


7. Social Robots and Active Appliances

This is the fourth pair of design metaphors.

Social Robots

Social robots are designed to interact with humans using human-like behaviors.

Examples include humanoid robots and companion robots.

However, social robots have limited commercial success due to complexity and cost.

Active Appliances
Active appliances are intelligent machines designed to perform useful tasks.

HARSH VERMA 9
Human-Centered AI (BAI613)

Examples include:

 Robotic vacuum cleaners

 Smart thermostats

 Smart home devices


These appliances are highly successful because they focus on functionality rather than human
imitation.

HARSH VERMA 10

You might also like