Audit trail

Type of pattern

• Process & product

 

Type of objective

• Trust

 

Target users

• Operator

 

Impacted users

• Employees
• Management
• Consumers
• Regulators
• Specific policy communities

 

Lifecycle stages

• Operation

 

Relevant AI ethics principles

• Privacy & security
• Reliability & safety
• Transparency and explainability
• Accountability
• Contestability
• Fairness
• Human-centred value
• HSE well-being

 

Benefits

• Enable accountability and regulatory conformance

 

Drawbacks

• Performance penalty
• Increased development effort

 

Known uses

• Falco, Gregory, et al. “Governing AI safety through independent audits.” Nature Machine Intelligence 3.7 (2021): 566-571.
• Winfield, Alan FT, and Marina Jirotka. “The case for an ethical black box.” Annual Conference Towards Autonomous Robotic Systems. Springer, Cham, 2017.
• Falco, Gregory, and Joshua E. Siegel. “A Distributed `Black Box’ Audit Trail Design Specification for Connected and Automated Vehicle Data and Software Assurance.” arXiv preprint arXiv:2002.02780 (2020).