Groups of SAIs and individual SAIs are producing interesting publications that are highly relevant for performance auditors. We recommend these:
Auditing Machine Learning Algorithms – a white paper for public auditors
By the Supreme Audit Institutions of Finland, Germany, the Netherlands, Norway and the UK (Published 24 November 2020)
The use of artificial intelligence is rising fast in governments across the globe. This will inevitably become an important field also for performance auditors. An informal group of SAIs recently produced a whitepaper on “Auditing machine learning algorithms”. This should be a highly useful guideline for performance auditors examining areas where machine learning or artificial intelligence is used. Four of the five SAIs involved in producing the guideline are PAS members, and the former chair of PAS, Mr Jan Roar Beckstrøm, has also been personally involved. You get access to the whitepaper on: auditingalgorithms.net
The whitepaper has been developed outside of INTOSAI´s due process for pronouncements, and is as such not part of the IFPP.
By SAI Netherlands
A report from the Netherlands Court of Audit (PAS-member) on “Understanding Algorithms” is also available in English: Rekenkamer_NL_understanding-algorithms
The Netherlands Court of Audit has investigated the Dutch government’s use of algorithms and developed an audit framework for this, which also considers ethical aspects of algorithms. The Court of Audit recommends that more account should be taken of questions from concerned citizens about the use of algorithms.
By Mr Jan Roar Beckstrøm, Office of the Auditor General of Norway
“Democracy demands that public auditors step up to provide assurance and accountability over governments’ use of algorithms, says Jan Roar Beckstrøm, deputy director general and chief data scientist in Norway’s national audit body”. In: Public Finance Focus – Viewpoints 30 March 2021