

#173 Building Trustworthy AI with Alexandra Ebert, Chief Trust Officer at MOSTLY AI
Jan 15, 2024
Alexandra Ebert, Chief Trust Officer at MOSTLY AI and a data privacy expert, delves into the challenges of building trustworthy AI. She discusses the critical need for ethical practices and transparency to regain public trust, highlighting the risks of bias and misinformation in AI systems. Alexandra emphasizes the role of synthetic data in improving accessibility and privacy while addressing fairness in AI outputs. The conversation also touches on the importance of user education regarding AI's limitations and the necessity for skilled professionals to navigate this complex landscape.
AI Snips
Chapters
Transcript
Episode notes
Trust Breaches in AI
- Apple's credit card faced accusations of discriminating against women, impacting user trust.
- Google's photo labeling AI mislabeled Black people, causing reputational damage.
Patterns of AI Mistrust
- AI failures cause PR disasters and harm users, leading to trust issues.
- Common reasons for losing trust include bias, discrimination, privacy breaches, and lack of explainability.
Building Trust in AI
- Some have overinflated expectations of AI's capabilities, while others are too mistrusting.
- Building trust and using AI cautiously in appropriate areas is an ongoing process.