A Public Administration Route to Algorithmic Transparency III: Thresholds for Transparency in the Private Sector
A regulatory focus on public law leaves out the bulk of algorithmic decision-making that we encounter in sectors such as finance, private healthcare, insurance, entertainment, and a host of other services. However, the features of administrative law which render it suitable for public sector use of
Vue d’ensemble
This is the third article in a series of three essays by researcher and Senior Trustworthy AI Fellow at Mozilla, Amber Sinha, about the domain of administrative law and how it is uniquely suited to guide the defining of thresholds for meaningful algorithmic transparency, both in public and private law.
In it, Sinha looks at principles in administrative law: duty to give reasons and administrative discretion as guiding principles for transparency in private law, focusing specifically on the provisions (Articles 13 and 14) of the EU's AI Act. The duty to give reasons is instructive to look at while arriving at thresholds for 'interpretability' under Article 13. Similarly, the requirement for a human agent to apply their mind while exercising administrative discretion is helpful in establishing the contours of human supervision, as envisaged under Article 14.
When high-risk systems employ algorithmic systems, their deployment must be preceded by the following conditions being met:
1.) The algorithmic systems must be designed to be interpretable to users such that the high-risk system may discharge the duty to give reasons for its decisions, as understood in administrative law.
2.) Where the technical nature of the algorithmic system poses fundamental interpretability challenges, it needs to be designed to flag sufficient information for independent human assessment to verify the machine’s inferences. Sufficient information may include the input data, the nature of the model in use and the likely factors which informed the decision.
3.) The algorithmic system must be designed for supervision by a human agent within the high-risk system.
4.) The nature of transparency afforded to the human agent in (c) must permit them to independently verify or validate decisions made by the algorithmic system.
