Lecture 1
Theories link concepts in order to explain something. Theories describe, explain and predict.
Theories reduce the complexity of reality. Theories can be modelled.
Design: shape, form, colour, functionality, affordances. The individual perceives the design. There
are individual differences (like gender, age, personality traits). User processing robot: attention,
perception, feeling, user psychology. How are robots perceived? Here is no uniform robot effect (no
“one-for-all” robot or best robot. The same robot can have different effects. Understanding
individual difference effects moderating and mediating variables.
Development, application and use of robots.
- Ambiguities in design
o The ‘thing’ vs the ‘perception’, affordances
- Ambiguities in language and understanding (miscommunication)
o Referencing
- Complexities in modelling
o Dynamic, multimodal
- Ethical aspects
o Privacy
Applications of robots
- Healthcare Paro
o Elderly
o Autism
- Education Nao
o STEM
o 2nd language
- Service profession Pepper
o Shopping mall
o Reception
Interaction with robots:
- Little emphasis is placed on the psychological construction of robots and testing them on
humans
- Understanding human behaviour
- Building theoretical models
- Conducting social science research
- Constructing valid and reliable measures
- Development of robots in socially acceptable ways
- Meet peoples expectations of its behaviour
Broadbent, pay attention on the following concepts:
- The Uncanny Valley
,The more lifelike a creation, the more likely it crosses the line from cute to creepy. Questions are
being raised as to its validity (whether this is really true). Possible explanations: evolution (death,
zombies: close to human but not totally human), expectation violation (perceive human like
expect to act human like)
- The Second Self
Robots Change, Challenge and Reveal Us (Turkle: Computer not as a “tool”, but as part of our social
and psychological lives). Reflects back to who we are and how we act in society.
The insights people gain about themselves through the use of computer (Then, Mobile phones, now
social robots).
Robots make us more self-reflective: questions what is it to be human (what is humanness, how do
we differ from robots?)
- The Media Equation
Computers Are Social Actors (CASA). People treat computers as if they were human even though
they know this is not the case. Do we ‘mindlessly’ apply social rules to robots? This is tested in many
studies:
People use stereotypical social categories (gender, ethnicity, in-/out-group status). Overlearned
behaviours (politeness; reciprocal, self-disclosure). Apply social rules ‘mindlessly’ (racial prejudice).
Also counter-evidence (not mindless).
- Anthropomorphism
Anthropomorphism is not Humanlike design. It is user’s experience. Stimulate such user experience
by Anthropomorphism in design = adding humanlike features. For example having the animal or
object behave as if it is human (La vache qui rit), Disney films. Anthropomorphism refers to
individual perception but is mostly conceived as a design feature.
Anthropomorphism refers to individual perception but is mostly conceived as a design feature.
Do we perceive the world through a human filter? Less predictable agents are anthropomorphized
to a greater extent. Inherent need to be social; less social ties, more anthropomorphism
- Mind Perception
Do we perceive mind in robots? 2 sub questions: Can robots think? And Do we think that robots
think. Theory of mind: perspective taking and understanding.
Perceived mind in two dimensions:
, - Capacity for agency (self-control, morality, emotion recognition…)
- Capacity for experience (hunger, pain, desire, personality, joy…)
More in lecture 5
- Relating to Robots
Studies observed emotional attachment to robots:
- E.g., measured by positive interactions
- Affective responses
- Hugging, smiling, caressing, sharing cookies
- Feels like a friend
- Feel empathy when robot is harmed
How and why connect to a mechanical friend?
- Physical Embodiment
Robots versus other technologies
- We relate differently to physically embodied robots than to virtual robot characters
(presented as avatars on a screen) and to computers…..
- Robot Abuse
Unguarded (groups of) children and youngsters tend to abuse a robot (in shopping malls, kicking
hitting, blocking path and bullying). Worries for bad social skills. To teach children that beating up a
robot is a bad idea: Shelly.
……
Lecture 2
Reverse engineering: interact with a robot and write pseudo code for (the improvement of) the
interaction
Scenario development and coding: Act out an interaction and anticipate error- human, hardware
and software
Omankwu, Nwagu and Inyiama: robotics and AI: differences and
similarities
Differences and Similarities with Common Computer-based AI
Robotics: the analogue world. Branch of electro-mechanical engineering. Robots are programmable,
physical machines, able to carry out a series of actions
Artificial Intelligence (AI): The digital world, branch of computer science. Computer programs that
complete tasks that otherwise require human intelligence. Robotics are not the same as AI
1. Robots interact with the physical world (with sensors, actuators)
2. Robots are programmable (similar to AI)
3. Robots are usually autonomous or semi-autonomous but do not have to be (cf. remote
control)
4. AI algorithms can tackle learning, perception, problem-solving, language-understanding
and/or logical reasoning -without sensors or actuators
, Combine those 2: an Artificially Intelligent robot.
AI is a part of robotics. Robotics also includes electro-mechanics and conventional programming.
This implies: requirement list, system analysis, system design, maintenance, updates, deployment,
testing and code design.
Webcrawler is not a robot even though we call it a robot. Software robot is NOT a physical robot,
only exists within a computer. AI runs on a host device, usually not standalone. Autonomously
completes a virtual task.
Use a simulated (virtual) agent to test software for your robot. IT IS NOT A ROBOT. The interaction
between virtual agents and human are different from interaction between humans and robots. You
can try it for you first trials etc. but not actually how people will respond to it irl.
Schneiderman (2020)
Common goals in AI research:
1. Theory development: Understand human/animal/plant abilities to build systems that
perform as well as or better than organisms
2. Applications: AI methods to built widely used products and services
Theory development in AI concerns:
- Patter recognition (images, speech, facial, signal, etc.)
- Natural language processing
- Translation of natural language
- Bipedal robots
- Emotionally responsive human faces
- Game playing (checkers, chess, go)
AI emulation methods often fail where conventional engineering solutions succeed. In 1997, IBM
Deep Blue chess-playing program defeated world champion. Feng-hsiung Hsu stated that brute force
hardware did the trick and not AI. In business, AI expert systems often failed where systems with
human-curated rule sets succeeded. This is because there is a finite search space.
Human emulation and simulation will offer worse solutions than ‘non-natural’ invention.
Mumford (1934)
- First attempts at new technologies are misled by human and animal models
- Four wheels have large advantages over two feet
- Airplanes have wings, but they do not flap like bird wings
The most ineffective kind of machine is the realistic mechanical imitation of a man or other animal
(for example let an airplane fly like a bird will NOT work!)
Use case is a description of a system’s behaviour, which responds to a request from outside the
system and describing “who” can do “what” with a system.
Summary:
Robot are electro-mechanical. They can have AI inside but not necessarily. Autonomy is no
prerequisite for robots (e.g. teleoperation). AI does not need sensors and actuators. Thus, in our
course, any AI that, apart form a computer, screen, keyboard, is not concerned with physical I/O
Theories link concepts in order to explain something. Theories describe, explain and predict.
Theories reduce the complexity of reality. Theories can be modelled.
Design: shape, form, colour, functionality, affordances. The individual perceives the design. There
are individual differences (like gender, age, personality traits). User processing robot: attention,
perception, feeling, user psychology. How are robots perceived? Here is no uniform robot effect (no
“one-for-all” robot or best robot. The same robot can have different effects. Understanding
individual difference effects moderating and mediating variables.
Development, application and use of robots.
- Ambiguities in design
o The ‘thing’ vs the ‘perception’, affordances
- Ambiguities in language and understanding (miscommunication)
o Referencing
- Complexities in modelling
o Dynamic, multimodal
- Ethical aspects
o Privacy
Applications of robots
- Healthcare Paro
o Elderly
o Autism
- Education Nao
o STEM
o 2nd language
- Service profession Pepper
o Shopping mall
o Reception
Interaction with robots:
- Little emphasis is placed on the psychological construction of robots and testing them on
humans
- Understanding human behaviour
- Building theoretical models
- Conducting social science research
- Constructing valid and reliable measures
- Development of robots in socially acceptable ways
- Meet peoples expectations of its behaviour
Broadbent, pay attention on the following concepts:
- The Uncanny Valley
,The more lifelike a creation, the more likely it crosses the line from cute to creepy. Questions are
being raised as to its validity (whether this is really true). Possible explanations: evolution (death,
zombies: close to human but not totally human), expectation violation (perceive human like
expect to act human like)
- The Second Self
Robots Change, Challenge and Reveal Us (Turkle: Computer not as a “tool”, but as part of our social
and psychological lives). Reflects back to who we are and how we act in society.
The insights people gain about themselves through the use of computer (Then, Mobile phones, now
social robots).
Robots make us more self-reflective: questions what is it to be human (what is humanness, how do
we differ from robots?)
- The Media Equation
Computers Are Social Actors (CASA). People treat computers as if they were human even though
they know this is not the case. Do we ‘mindlessly’ apply social rules to robots? This is tested in many
studies:
People use stereotypical social categories (gender, ethnicity, in-/out-group status). Overlearned
behaviours (politeness; reciprocal, self-disclosure). Apply social rules ‘mindlessly’ (racial prejudice).
Also counter-evidence (not mindless).
- Anthropomorphism
Anthropomorphism is not Humanlike design. It is user’s experience. Stimulate such user experience
by Anthropomorphism in design = adding humanlike features. For example having the animal or
object behave as if it is human (La vache qui rit), Disney films. Anthropomorphism refers to
individual perception but is mostly conceived as a design feature.
Anthropomorphism refers to individual perception but is mostly conceived as a design feature.
Do we perceive the world through a human filter? Less predictable agents are anthropomorphized
to a greater extent. Inherent need to be social; less social ties, more anthropomorphism
- Mind Perception
Do we perceive mind in robots? 2 sub questions: Can robots think? And Do we think that robots
think. Theory of mind: perspective taking and understanding.
Perceived mind in two dimensions:
, - Capacity for agency (self-control, morality, emotion recognition…)
- Capacity for experience (hunger, pain, desire, personality, joy…)
More in lecture 5
- Relating to Robots
Studies observed emotional attachment to robots:
- E.g., measured by positive interactions
- Affective responses
- Hugging, smiling, caressing, sharing cookies
- Feels like a friend
- Feel empathy when robot is harmed
How and why connect to a mechanical friend?
- Physical Embodiment
Robots versus other technologies
- We relate differently to physically embodied robots than to virtual robot characters
(presented as avatars on a screen) and to computers…..
- Robot Abuse
Unguarded (groups of) children and youngsters tend to abuse a robot (in shopping malls, kicking
hitting, blocking path and bullying). Worries for bad social skills. To teach children that beating up a
robot is a bad idea: Shelly.
……
Lecture 2
Reverse engineering: interact with a robot and write pseudo code for (the improvement of) the
interaction
Scenario development and coding: Act out an interaction and anticipate error- human, hardware
and software
Omankwu, Nwagu and Inyiama: robotics and AI: differences and
similarities
Differences and Similarities with Common Computer-based AI
Robotics: the analogue world. Branch of electro-mechanical engineering. Robots are programmable,
physical machines, able to carry out a series of actions
Artificial Intelligence (AI): The digital world, branch of computer science. Computer programs that
complete tasks that otherwise require human intelligence. Robotics are not the same as AI
1. Robots interact with the physical world (with sensors, actuators)
2. Robots are programmable (similar to AI)
3. Robots are usually autonomous or semi-autonomous but do not have to be (cf. remote
control)
4. AI algorithms can tackle learning, perception, problem-solving, language-understanding
and/or logical reasoning -without sensors or actuators
, Combine those 2: an Artificially Intelligent robot.
AI is a part of robotics. Robotics also includes electro-mechanics and conventional programming.
This implies: requirement list, system analysis, system design, maintenance, updates, deployment,
testing and code design.
Webcrawler is not a robot even though we call it a robot. Software robot is NOT a physical robot,
only exists within a computer. AI runs on a host device, usually not standalone. Autonomously
completes a virtual task.
Use a simulated (virtual) agent to test software for your robot. IT IS NOT A ROBOT. The interaction
between virtual agents and human are different from interaction between humans and robots. You
can try it for you first trials etc. but not actually how people will respond to it irl.
Schneiderman (2020)
Common goals in AI research:
1. Theory development: Understand human/animal/plant abilities to build systems that
perform as well as or better than organisms
2. Applications: AI methods to built widely used products and services
Theory development in AI concerns:
- Patter recognition (images, speech, facial, signal, etc.)
- Natural language processing
- Translation of natural language
- Bipedal robots
- Emotionally responsive human faces
- Game playing (checkers, chess, go)
AI emulation methods often fail where conventional engineering solutions succeed. In 1997, IBM
Deep Blue chess-playing program defeated world champion. Feng-hsiung Hsu stated that brute force
hardware did the trick and not AI. In business, AI expert systems often failed where systems with
human-curated rule sets succeeded. This is because there is a finite search space.
Human emulation and simulation will offer worse solutions than ‘non-natural’ invention.
Mumford (1934)
- First attempts at new technologies are misled by human and animal models
- Four wheels have large advantages over two feet
- Airplanes have wings, but they do not flap like bird wings
The most ineffective kind of machine is the realistic mechanical imitation of a man or other animal
(for example let an airplane fly like a bird will NOT work!)
Use case is a description of a system’s behaviour, which responds to a request from outside the
system and describing “who” can do “what” with a system.
Summary:
Robot are electro-mechanical. They can have AI inside but not necessarily. Autonomy is no
prerequisite for robots (e.g. teleoperation). AI does not need sensors and actuators. Thus, in our
course, any AI that, apart form a computer, screen, keyboard, is not concerned with physical I/O
