Published: 05 September 2022
Summary
Training, calibrating, validating and explaining AI-enabled systems requires human-in-the-loop architecture and user interfaces. Data and analytics leaders must enable this relationship between humans and algorithms to improve speed, accuracy and risk compliance.
Included in Full Research
Overview
Key Findings
Human-in-the-loop (HITL) applications and interfaces are needed to interact with a wider set of stakeholders than just data scientists. Society, business and individuals must be able to interface with artificial intelligence (AI) to shape the behavior of autonomous systems for performance improvement and risk mitigation.
As AI systems’ development and deployment increases, the roles and tasks of staff change as they inherit new responsibilities within HITL AI systems.
Bridging human understanding and communication with opaque AI systems is a huge cognitive design challenge and often a roadblock to wider user acceptance and adoption.
Recommendations
For data and analytics leaders looking to scale
To view the entire document, log
in or purchase