Skip to Content

Wiki source code of b. Design Scenario

Version 3.2 by Michaël Grauwde on 2023/05/07 12:26
Hide last authors
Michaël Grauwde 2.1 1 __Conceptual Scenario__
2
Michaël Grauwde 3.1 3 Our conceptual scenario revolves around our member of the public attempting to communicate with the deliberative agent to reflect on their values.
Michaël Grauwde 2.1 4
Michaël Grauwde 3.2 5 **Example: **Helen gets an invitation to join a workshop given by TU Delft, TNO, and the police. She knows only that the workshop is about if AI systems should be developed for public safety to be used by the police. She has been hearing about this AI thing and sees that it's all the hype and realises that it is probably sooner rather than later going to be used by the police, so she wants to have a say in how it will be used.
Michaël Grauwde 2.1 6
Michaël Grauwde 3.2 7 As she walks into the workshop room, she is brought to a computer by one of the employees there. The employee tells her that she is going to have a conversation with a chatbot and based on that conversation, the chatbot will aim to see which values are important to her. Slightly scared, Helen gets behind the laptop. As she looks at the computer, the chatbot opens up.
Michaël Grauwde 2.1 8
Michaël Grauwde 3.1 9 **Dialogue**
Michaël Grauwde 2.1 10
Michaël Grauwde 3.1 11 CA: "Hi, how are you today?"
Michaël Grauwde 2.1 12
Michaël Grauwde 3.2 13 H: "I'm good, thank you very much. What are we doing today?"
Michaël Grauwde 2.1 14
Michaël Grauwde 3.1 15 CA: "Today we are going to engage in a serious of discussions to see what values are important to you when it comes to public safety"
Michaël Grauwde 2.1 16
Michaël Grauwde 3.2 17 H: "That is a bit scary but i guess i'm already here."
Michaël Grauwde 2.1 18
Michaël Grauwde 3.1 19 CA: "There's no reason to be scared. My goal is to communicate your values back to you and help you reflect on them so that you can engage in better deliberations with your fellow humans."
Michaël Grauwde 2.1 20
Michaël Grauwde 3.2 21 H: "Ok, i guess"
Michaël Grauwde 2.1 22
Michaël Grauwde 3.1 23 CA: "Perfect! Let us Start!"
24
Michaël Grauwde 3.2 25 Helen is presented with with various questions about what is important to her.
Michaël Grauwde 3.1 26
27 After some interactions, the chatbot says,
28
Michaël Grauwde 3.2 29 CA: "thanks for answering these questions Helen! We have determined that //freedom //is very important for you."
Michaël Grauwde 3.1 30
Michaël Grauwde 3.2 31 H: "Thank you!"
Michaël Grauwde 3.1 32
33 CA: "Will you participate in another exercise based on a scenario"
34
Michaël Grauwde 3.2 35 Helen participates in another scenario that is based on a crowd safety scenario.
Michaël Grauwde 3.1 36
37 In this scenario, she gets a value that stays at //freedom//.
38
39 The conversational agent states,
40
41 CA: "You still have the value of //freedom//. Now we would like you collaborate deliberate with the rest of the humans in the room about which values they received."
42
Michaël Grauwde 3.2 43 Helen deliberates with her fellow humans in the room and comes away with a better reflection of why others feel how they do about //security//, //responsibility,// and //autonomy//.