"Transparency International is concerned specifically with the corrupt use of AI, defined as the intentional abuse of AI systems for the private benefit of entrusted powerholders. This is distinct from issues arising from unintended consequences and outcomes of the adoption of AI systems. In particular, this working paper covers the corrupt use of AI in the public sector, which can take many forms. For example, in resource allocation, manipulated AI systems might disproportionately direct funds to particular regions or political constituencies. In policy implementation, AI-driven models could be skewed to support predetermined outcomes, undermining fair and evidence-based policymaking. In public procurement, AI systems could be intentionally programmed to unfairly prioritise certain vendors. Generative AI can be abused by corrupt actors to fuel disinformation campaigns that undermine political integrity. AI can also be used to enable corrupt actors to amplify disinformation campaigns, manipulate public discourse, and enhance censorship and surveillance, thereby undermining civil and political rights. These abuses of AI intensify existing threats to transparency and fairness in governance, particularly in the public sector. Additionally, while many governmental, non-governmental and private-sector organisations are active in the AI integrity and ethics space, there is currently no deliberate focus on, or call for, measures that ensure AI is not used for corrupt outcomes or to facilitate corrupt acts. At the policy level, the current discourse generally overlooks the intentional misuse of AI. This working paper therefore sheds light on how corruption relates to AI integrity, by defining the corrupt use of AI, providing overviews of the main manifestations of such use, and making recommendations on how to address these challenges." (Page 3)