Blog Revolutionising anti-bribery and corruption procedures through Artificial Intelligence

The opinions expressed here are those of the authors. They do not necessarily reflect the views or positions of UK Finance or its members.

Artificial intelligence ("AI") is rapidly transforming the anti-financial crime landscape globally, including anti-bribery and corruption ("ABC") procedures. Below, we explore the real-life use-cases for AI within ABC programmes and set out the next steps for financial institutions ("FIs").

Pioneering use-cases of AI in ABC across the financial sector

A prominent use-case for AI lies in transaction monitoring ("TM"). In a high-profile collaboration, HSBC partneredwith Google to develop a machine learning ("ML")-enabled system to help flag criminal activities. They are now reportedly intercepting 50 per cent more suspicious transactions, while finding two to four times more true hits. With irregular payment patterns and hidden connections better uncovered, enhanced ABC efficiency and accuracy enables HSBC to focus resources on significant threats.

Another key element of ABC, third-party due diligence ("TPDD"), is transformed by automated risk assessments. Leveraging structured global data, AI agents scan for red flags and expose third-party relationships in real time. These innovations promote interventions and targeted enhanced due diligence. Standard Chartered integratedAI tools, powered by natural language processing ("NLP"), in the screening of vendors against sanctions and adverse media, reducing false positives by 25 per cent and speeding up TPDD.

AI can also proactively manage compliance risks. It enhances horizon-scanning by continuous analysis of ABC regulatory updates across jurisdictions, ensuring FIs stay ahead of requirements. Tools using Generative AI ("GenAI") and big data analytics track changes in laws, predict compliance gaps, and extract new rules for FIs. They can also automate ABC procedural updates, facilitating swift adaptation to evolving landscapes and enabling compliance teams to prioritise high-risk areas.

Meeting challenges arising from AI and the next steps FIs can take

AI integration in ABC does not come without complications. For instance, algorithmic bias may unfairly flag certain sociodemographic groups; opaque AI-generated decisions may impede accountability; and abuse by corrupt actors to exploit system vulnerabilities may amplify rather than mitigate risks.

These challenges necessitate robust governance and guardrails. Here are four recommendations for effective and ethical AI deployment:

1. Suitability assessment is a prerequisite for matching an AI solution to the target problem. This involves evaluating data availability, task complexity, and organisational resources. For example, ABC algorithms performing well with large datasets may struggle in corruption hotspots known for low digitisation rate, hindered by the lack of high-quality input.
2. Adopting Explainable AI ("XAI") with human-in-the-loop integrates expertise into automated processes. This augments XAI decisions in ABC with transparency and verifiability, allowing oversight and mitigating errors. SMEs must be regularly engaged to test and finetune applications. This is essential for accountability and collaboration with regulators.
3. Benchmarking best practices is key to aligning AI strategies against peer standards. Investing in staff training on AI and ABC, consulting with global leaders in the industry, and fostering interaction with SMEs, are vital to ensure AI-related improvements are rolled out sustainably.
4. Model assurance allows FIs to audit the AI system such that it performs as expected, without algorithmic bias and significant vulnerabilities, while also helping assure against model drift. This final guardrail ensures that AI applications are secure, impartial, and trustworthy.

The trend of AI transformation in ABC will now only accelerate. Agentic AI, with its ability to plan and execute tasks autonomously, could be a further leap. FIs should capitalise on not only industry-leading examples, but also the above ethical measures. With a holistic approach to AI, FIs will bring the global fight against bribery and corruption to the next level.