This week we interview Irio Musskopf (@irio), one of the founders of the AI-based anti-corruption project Operacao Serenata do Amor (https://serenata.ai/en/) in Brazil. The interview provides detailed insights into how open government data paired with intelligent algorithms can be used to promote integrity.

The interview kicks off with a short outline how the panel on AI as an Anti-corruption tool chaired by Nils, Christopher and Steven Gawthorpe (@SteveGawthorpe) at last year’s Interdisciplinary Corruption Research Forum in Kyiv (https://www.icrnetwork.org/what-we-do/conferences/icr-forum-kyiv-2019/) and a recent report by Per Aarvik for U4 (https://www.u4.no/publications/artificial-intelligence-a-promising-anti-corruption-tool-in-development-settings) inspired Kickback to invite Irio for the interview.

In the interview, Irio outlines in much detail how the project came about, what challenges the team had to overcome and how it has been received by journalists, politicians and the public.

Irio describes the statistical approach used to detect suspicious spending patterns, referring to normal distribution and standard deviations. In case statistics is not your strong suit, you can check out this short explanation (https://statisticsbyjim.com/basics/normal-distribution/)

Chris, Nils and Irio also discuss the role that machines and humans play in such new anti-corruption efforts, whether and where humans might be replaced by intelligent algorithms and which tasks require the involvement of human decision making.

The interview covers the Twitter bot Rosie da Serenata (@RosieDaSerenata) that automatically tweets out suspicious cases.

For some further reading:

* Report published at the end of the 3 months financed by the first crowdfunding campaign.

https://medium.com/serenata/the-last-serenade-65fc1a9a0e2f

* Measuring the impact of the operation after one year of the first mass report to the Congress

https://medium.com/serenata/o-impacto-do-controle-social-na-c%C3%A2mara-dos-deputados-c2b2a34db09e