From Human-AI Interaction to UX Design Framework
Trust in machines is a new concept for many, unlike the trust we place in other humans. By replicating human trust mechanisms, we can enhance our relationship with AI. Trust can be divided into two components: cognitive and affective.
- Cognitive Trust : Cognitive trust is based on skills, reliability, and accomplishment. It is evaluated through tasks, qualifications, and interactions and originates from the brain and mind.
- Affective Trust : Affective trust focuses on likability, appearance, and non-verbal cues. It is judged by tone of voice, body language, and word choice, deriving from emotions and feelings.
By considering both cognitive skills and affective elements, we can establish a balanced approach to trust in technology, making trusting machines a more familiar process. Ensuring that users retain control and understand machine recommendations is key to fostering a positive relationship between humans and AI.
Step 1 : Onboarding to AI
Effective AI design begins before users engage with the system. It’s crucial to explain the system's capabilities and limitations upfront to set proper user expectations and build trust. Allowing users to test the system provides a sense of control, and transparency about privacy, security, and permissions is essential for establishing trust.
Step 2 : Building cognitive trust
Explainability and controllability are significant in AI interactions. AI systems should explain their decisions to users, similar to how humans seek justifications. This builds trust in the machine's judgment and fosters comfort with its decisions. Users must have the ability to intervene and influence AI outcomes based on their preferences.
Transparency and user control are imperative for a positive user experience. For instance, an AI application that allows users to control generated results enhances trust and satisfaction.
Step 3 : Designing AI with affective trust
Establishing an emotional connection with users is critical. Visual elements like colors, icons, and data visualizations play a significant role in building trust. Studies, such as those conducted by Harvard University on a medical diagnosing system, show that the tone of voice and task flow significantly influence users' trust in the system's recommendations.
Using a humble tone of voice and speaking in the user's language are vital to conveying the AI’s awareness of user needs and its capabilities, fostering trust. The right color schemes, like blue for trust in social networks, and visual elements like icons and progress bars enhance comprehension and trust.
Step 4 : Celebrating success with AI
After a task is completed, highlighting the efficiency of results is crucial. This can be done through data visualization or simple text, demonstrating the benefits of AI over manual work. Respecting user preferences during tasks and showing that these preferences are acknowledged fosters cognitive trust.
For example, if a user requests only the top three insights, this preference should be respected and reflected in future interactions, demonstrating the AI system's responsiveness.
Usecase : Using AI as a UX framework in context
In a project at Ericsson called "Zero Touch Network," efforts were made to automate network infrastructure and orchestration. Clear expectations were set for engineers about the system’s capabilities and limitations, and decisions made by the AI system were explained in understandable terms. A feedback mechanism allowed engineers to reject system recommendations and explain their reasons, improving the system’s adaptability.
Discover this tale amongst others in Ahmed's Story
