AI ethics risk management
Over the past years the use of machine learning (ML) in the financial services sector has become ubiquitous.
As the financial industry is built on data this is unsurprising. AI technologies can lead to better forecasting, pricing models, more targeted products and automation of back-office operations.
AI technologies raise both novel challenges, such as ethical issues related to AI bias, and deepen existing risks to consumer, financial soundness of firms and systemic financial stability.
In partnership with Holistic AI, we provide AI risk management services to help firms prevent, manage and monitor AI risks as they materialise across the business and beyond.
AI ethics risk impact assessment
Intification of AI capabilities in product lifecycle, including in house and any external models
Analysis of existing documentation (e.g. policies, internal wikis) and code
Interview with developers and product managers
Assessment of AI ethics risk across the following dimensions:
§ Bias
§ Explainability
§ Privacy
§ Robustness
Mix of qualitative assessment via questionnaires with developers and product managers and technical code audit review
Review of findings from initial analysis against industry benchmarks and regulatory expectations
AI ethics risk management framework
Design policy outlining AI ethics governance (included in product design and ongoing review)
Senior management responsibility allocation and oversight procedures
Design AI ethics risk management: risk identification, risk impact rating, risk appetite thresholds, management information, ongoing risk monitoring efficacy methodology
Design of relevant controls at code, business process and training & awareness level to ensure AI ethics risks identified are appropriately managed
Deliver face to face or online module training to relevant staff about the risks of AI and ways to prevent and manage AI ethics liability
Design of ongoing assurance mechanisms to test efficacy of risk management and control processes