Trustworthy AI in practice: an analysis of practitioners' needs and challenges

Existing frameworks aren't being widely used in practice: Despite the proliferation of Trustworthy AI frameworks and guidelines in recent years, the research reveals a gap between what academics have created and what practitioners actually employ when developing AI systems, suggesting current resources may not meet real-world needs.

Practitioners face significant practical challenges implementing ethical AI principles: The study identifies concrete obstacles that AI developers encounter when trying to incorporate trustworthiness principles into their work, indicating that knowing what to do and actually doing it are two different problems requiring targeted support.

There's a mismatch between available resources and what practitioners need: Through surveys and interviews, the research found that practitioners lack adequate tools, knowledge, and guidelines specifically designed for their implementation challenges, suggesting that current academic recommendations may be too theoretical or disconnected from development workflows.

Practitioners have their own vision of what Trustworthy AI means: The study reveals that how practitioners understand and prioritize Trustworthy AI principles differs from academic frameworks, highlighting the importance of grounding responsible AI guidance in practitioners' actual experiences and contexts.

Targeted recommendations can bridge the gap: The research provides specific recommendations to help AI practitioners better develop Trustworthy AI applications, suggesting that solutions exist but require better alignment between academic guidance and practical implementation realities.