Process Log 4

Date

Process Log 4

9/17-9/24

What is Responsible AI?

CONTEXT

What is Responsible, Explainable AI? After reading the definition of "interface" It is the connection between the seen and unseen. What are the information that is hidden behind the one interface?

Lack of transparency in using AI leads to easily overusing AI without critical thinking.

How might designing "ethical nudges" change behavior and perception of using AI in work - Ethical Nudges that are designed to surface three hidden harms: algorithmic bias, environmental cost, and misinformation risk. Making the Hidden Visible: "Ethical Nudges" for Bias, Environmental Cost, and Misinformation in Intelligent Interfaces

1. What is a “nudge”?

A nudge comes from behavioral economics (Thaler & Sunstein, Nudge, 2008). It means subtly shaping people’s choices without restricting freedom—making the ethical or beneficial path the “easy” one. Example: putting fruit at eye level in a cafeteria instead of banning junk food.

2. What is an AI nudge?

An AI nudge uses artificial intelligence to deliver or adapt nudges dynamically. Instead of static defaults, AI can analyze user behavior, context, or preferences to personalize and optimize nudges in real time. Process

3. What is Responsible AI?

4. What is Transparent AI?

5. What is Explainable AI?

PROCESS

Research different software companies' AI Principles helped me provide the main terminologies related to AI ethics. Figma, Adobe, Google AI, Anthropic, Google Cloud, and Midjourney were selected to study.

REFLECTIONS

I am not sure if designers would really care. I am not sure if they really want to see this behind AI. How would I check their perception and emotion when they are looking at this information? Do these information really matter to them? Maybe it doesn't matter to them anymore as long as the feature is working for them.

INSPIRATION/NEW QUESTIONS

How do I show invisible to visible? What information should I show and not show? How much information should I show? Is this another cognitive burden? What kind of service do I want to test on?

CONCRETE ACTION

Literature Reviews on How nudge subtly changes behavior and perception.

1. Mapping the Ethics in AI (cost) to set up crierias (cost, energy, copyright, biased data, over reliance, well being)
2. Iterations of Interface Designs
3. Types of nudging (Timely? how often, mobile? web? what format, where? how?Tone of message?)
4. Create a survey of AI usage preconception and post conception Questions can be, "How did their perception, emotion, perception of using AI changed?"
5. Short User Testing the users with the prototype.
- Behavior path
- What actions did change among participants
6. Long diary journal before and after to test how participants changed perception, emotion of using AI changed?
- Measure Time, frequency of using AI in their daily life