Ethical Machine Learning - Kehsa Williams - NDC Porto 2023

Discover the critical role of ethics in machine learning, including bias detection, transparency, and explainability, with a focus on human oversight, AI system performance, and societal implications.

Key takeaways

Importance of ethical considerations in machine learning
Data preprocessing and bias detection during the data prep stage
Explainability and transparency in AI decision-making
Need for human oversight and feedback in AI systems
Impact of AI on people and society, such as biased decision-making and lack of explainability
Importance of understanding and addressing bias in AI systems
Use of tools like Data Wrangler and SageMaker Clarify to detect and mitigate bias
Example of AI system used in police work, highlighting potential for bias and lack of transparency
Need for regular monitoring and reporting of AI system performance and results
Importance of considering moral and ethical implications of AI systems
Example of biased decision-making in hiring, advertising, and lending
Need for transparency and explainability in AI decision-making
Importance of considering both ethical and technical aspects of AI development
Use of AI to identify breast cancer early, but also potential misuse and biased decision-making
Need for continuous learning and improvement in AI development

Ethical Machine Learning - Kehsa Williams - NDC Porto 2023

More talks