Examining the Black Box. Tools for assessing algorithmic systems
Ada Lovelace Institute published a report to shed on algorithm audit and algorithmic impact assessment
They classified these tools in the following manner
1) Algorithm audit
They distinguish between algorithm audits (from the perspective of the computer science community) and algorithm inspections (from the perspective of its use in common language meaning broad inspection and compliance)
— Bias audit: a targeted, non-comprehensive approach focused on assessing algorithmic systems for bias.
— Regulatory inspection: a broad approach, focused on an algorithmic system’s compliance with regulation or norms, necessitating a number of different tools and methods; typically performed by regulators or auditing professionals.
2) Algorithmic impact assessment
— Algorithmic risk assessment: assessing possible societal impacts of an algorithmic system before the system is in use (with ongoing monitoring often advised), to assess the possible areas of impact of the system and the attendant risk.
— Algorithmic impact evaluation: assessing possible societal impacts of an algorithmic system on the users or population it affects after it is in use
H/T @DorotheaBaur (via Twitter)
0 comentarios