Summary 1JM50 - Implementing and
Adapting to Artificial Intelligence in
Organizations
Week 1
Lecture 1
Diversity of systems
o Depending on the type of system, the shared task and the collaborative context, the impact on workers’ jobs
can largely differ, as well as their reactions to it
o To improve implementation and use of AI systems in work environments, we need to carefully compare what
the intended collaboration looks like, and how humans are involved in the process
o Thus, introducing AI at work is complex and dynamic
Mixed reactions: AI at work creates mixed reactions
o Users often display a positivity bias regarding AI capabilities
Overreliance even if system suggestions are wrong
Expecting perfect system performance
o Show resistance to use highly autonomous systems (low levels of trust and user intention)
High workload, as tasks are not delegated to system
High reliance on manual decision making, even if this leads to safety risks or performance decrements
Perspectives from industry: Erno Ledder Data Science at Pogona Insurance
o Definition AI: Learning a model using statistics in order to allow a computer to make predictions or
classifications
o Challenges when developing AI
The challenge often depends on the use case
Difficult is to determine
What do I want to measure (and can my data be used for this)
What kind of data do I need and is my data biased etc.
What am I actually measuring
How to test confidence of the model
How to explain model/results
How to implement it in the final product
Data. Often the quality of data you work with is sub-optimal. ”Garbage in, garbage out”
o AI can be applied in numerous fields and is an inescapable part of the future. Therefore we need to make
sure that AI is developed correctly.
Ask questions, what kind of data was used, how was the model tested
, Keep an open mind, AI has way more applications than is depicted in the media
AI is not good or bad by nature, it is what you do with it. So use it for all the good it can do!
Video 1: Technology Acceptance Model
Technology Acceptance Model (TAM): why users may want to choose to work with the technology
o Some features may impact cognitive responses a user has towards the technology
Resulting in behavior: use the system or not
o Perceived usefulness: Degree to which a person believes that using a particular system would enhance their
performance
o Perceived ease of use: Degree to which a person believes that using a system is free from effort
o Users, systems, and context
Theory of Reasoned Action
o Behavior results from our attitude towards the behavior and subjective norms (would others engage in this
behavior)
Theory of Planned Behavior
o Adds the aspect of perceived behavioral control: heavily
impacts the decision on whether to engage in a behavior
Unified Theory of Acceptance and Use of Technology (UTAUT)
,Video 2 Organizational Change Management
Organizational Change: the planned alterations of organizational components (emergent changes are not
considered in this video)
o External and internal forces trigger organizations to change
Organizational Change Management: the process of guiding organizational change to a successful resolution.
LEWIN’S MODEL: three stages of change
1. Unfreezing: making people aware of the need for change and improving their motivation for change
Effective communication is key
2. Changing: change implementation
Careful planning, effective communication, and people involvement are necessary
3. Refreezing: moving employees from a state of change to a more stable state, accept and internalize the new
way of working
Requires positive reinforcement in terms of reward and recognition of employees
BECKHARD & HARRIS’ MODEL
1. Initial organizational analysis
2. Identifying need for change & vision
3. Gap analysis
4. Action planning
5. Managing the transition
Conclusion: Organizational Change Management is
dealing with the people side of organizational change.
It’s a process consisting of various consecutive stages. It
is crucial for the success of organizational change.
Video 3 Collaboration
Definition collaboration: “working jointly with others […] especially in an intellectual endeavor”
o Usually collaboration results from an individuals goal or desire solve/create/discover something. It involves
decision-making between different parties and a collective responsibility for an outcome.
Human collaboration: different aspects influence the effectiveness
o Communication (Speech, Gestures, Gaze, Nonverbal cues)
o Shared understanding
o Trust
o Context (Face to face vs. virtual collaboration)
o Individual characteristics
Personality (e.g., openness)
Experience
Knowledge and skills
Collaboration via technology: Effectiveness of virtual collaboration is dependent on technology related aspects
o User technology expertise
o Technology use challenges (e.g., error)
o Changes in how collaboration and communication takes place
, Collaboration with technology
o Similarities to human-human collaboration
o Many current challenges:
Communication
Trust
Acceptance of the technology
Understanding the impact of user and system characteristics
Video 4 Comparing Settings
Comparing AI Systems
o Levels of comparison
o Cooperation vs. Collaboration
o Comparing collaborative tasks, workspaces, and levels of collaboration
Human-AI interaction: systems can be compared in terms of three central factors regarding collaboration with
humans
1. Human characteristics
2. AI characteristics > IN THIS VIDEO
3. Physical environment they interact in (physical, virtual, or mixed)
Levels of Comparison
o System characteristics: Interface, Anthropomorphism, Functionalities
o Role and Task of the system: Level of Autonomy, task specifications
o Collaboration with humans: Role of human in interaction, proximity
Robot Taxonomy: when describing human-robot collaboration, it is important to consider what role the human
operator takes or how the human and the robot communicate
System Characteristics
o Embodied and virtual
o Anthropomorphism describes the tendency to imbue the real or imagined behavior of nonhuman agents
with humanlike characteristics, motivations, intentions, or emotions
Two motivational determinants
1) Experience confidence (sense-making)
2) Form social bonds
Taxonomy of collaborative tasks: The way a collaboration is set up will impact what kind of tasks the operator
and the system do as well as how much interaction takes place
Interaction to
collaboration