Week 6 - Solving Problems Flashcards
(7 cards)
Design Ethnography
Observing how users interact with the prototype in real settings, to uncover real-world challenges.
Earliest step in design process to avoid silly mistakes or find opportunities for innovation.
What assumptions does design ethnography question
- Is the environment different from where the device has been previously tested and deployed successfully?
- Which rules are flexible and open to change in this context?
- What are the motivations of people involved?
- Are you addressing the right problem or can you solve a more critical one?
- If you solved a related problem, could you eliminate this problem?
Participatory design
Actively involve users in refining the design, using their insights to minimise risks.
Using personas
Can use scenarios and personas to help identify risks by using assessing from multiple different points of view.
Risk mitigation checklist
- Zone 1 - Authority and discipline: will tech undermine authority, how could it affect children, could it allow children to escape just punishment, could it have restrictions/boundaries set, should it require supervision to be used.
- Zone 2 - Malevolent and accidental harm: how could it be used malevolently, how will malevolent use be monitored and prevented.
- Zone 3 - Emotionality and socialisation: does the tech appeal to emotionally vulnerable people, can it intensify this vulnerability, could the tech isolate people, what emotional state does the tech foster.
- Zone 4: Governance and accounting: if people will be producing extra content/components how will they be verified, what tools will a user be interacting with e.g., will a secure ecosystem be needed if server contact is frequent, is it clear where data is collected and stored.
Value-sensitive design
Integrating ethical and social values into the design process of a product.
Humble AI
An AI approach focused on ensuring trustworthy people are not unfairly distrusted, promoting fair, merit-based opportunities.
Challenging to balance fairness without increasing false positives.