Nils Köbis is a researcher based in Germany and works as a professor for the human understanding for algorithms and machines (1:16). Nils tells us how he came to work on the relationship between corruption and artificial intelligence (AI) (2:46). He then explains the origin of AI and what AI means (5:29). Nils shares examples where AI can be leveraged to tackle corruption (9:45). The type of training data provided to AI is important (15:04). For example, if the data is biased, the biases will be reproduced. In some cases, AI can be manipulated for corruption (18:37). Nils tested how humans react to ethical and non-ethical advice from AI and the results are surprising (22:45). He makes recommendations to limit the risks associated with AI and corruption (27:27).

To stay in touch with Nils:

• X (ex-Twitter): @NCKobis

Additional resources:

• Tweet Bot Rosie: @RosieDaSerenata
• Nils Köbis et al. 2022. The promise and perils of using artificial intelligence to fight corruption. Nils Köbis. 2023. Bribes for bias : can AI be corrupted?
• Transparency International. 2022. The corruption risks of artificial intelligence.
• Daniel Sejerøe Hausenkamph. 2023. The role of technology in anti-corruption: dystopian reflections.

Credits:

Sophie au pays des possibles is a podcast created, directed, and produced by Sophie Lemaître. This episode was recorded online in May 2024. Podcast illustration and music by Fabien Doulut. Editing by Manuel Perreux for the English version and Sophie Lemaître for the French version.