[AI Ethics Seminar EP.3] Introducing Key AI Toolkits for Designing Trustworthy AI Experiences and a Practical Application Guide - LG AI Research Blog

While generative AI technology is creating business value, limitations such as Hallucination and data bias can lead to corporate reputational damage and legal liability. Therefore, designing Human-AI Interaction with security and trustworthiness is critically important.

This article introduces various AI toolkits for maximizing trustworthy AI experiences.

• Microsoft's HAX (Human-AI Experience Toolkit): Provides guidelines and frameworks to help AI understand user needs and build trust. It covers a 4-stage interaction process and 18 guidelines, including solutions for when the system doesn't work properly.
• Google's PAIR (People + AI Research) Guidebook: Supports the design of AI systems that prioritize user experience.
• LG AI Research's AI Ethics Impact Assessment System: Introduces LG AI Research's own system, built based on reviews of various toolkits.

To apply Microsoft HAX to actual development, a systematic 5-step process is required, including selecting stakeholders, prioritizing guidelines, and verifying the expected application model.