{"id":128141,"date":"2023-11-13T14:10:46","date_gmt":"2023-11-13T14:10:46","guid":{"rendered":"https:\/\/feedzai.com\/?p=128141"},"modified":"2024-04-09T09:12:08","modified_gmt":"2024-04-09T09:12:08","slug":"enhancing-ai-model-risk-governance-with-feedzai","status":"publish","type":"post","link":"https:\/\/feedzai.com\/blog\/enhancing-ai-model-risk-governance-with-feedzai\/","title":{"rendered":"Enhancing AI Model Risk Governance with Feedzai"},"content":{"rendered":"
[vc_row row_height_percent=”0″ override_padding=”yes” h_padding=”2″ top_padding=”1″ bottom_padding=”2″ overlay_alpha=”50″ gutter_size=”3″ column_width_percent=”100″ shift_y=”0″ z_index=”0″][vc_column width=”1\/1″][vc_row_inner][vc_column_inner width=”1\/12″][\/vc_column_inner][vc_column_inner width=”10\/12″][vc_single_image media=”128142″ dynamic=”yes” media_width_percent=”100″ uncode_shortcode_id=”140946″][\/vc_column_inner][vc_column_inner width=”1\/12″][\/vc_column_inner][\/vc_row_inner][vc_row_inner][vc_column_inner column_width_percent=”100″ gutter_size=”3″ overlay_alpha=”50″ shift_x=”0″ shift_y=”0″ shift_y_down=”0″ z_index=”0″ medium_width=”0″ mobile_visibility=”yes” mobile_width=”0″ width=”2\/12″][\/vc_column_inner][vc_column_inner width=”8\/12″][vc_custom_heading heading_semantic=”h3″ text_size=”h3″ text_weight=”400″ uncode_shortcode_id=”177611″]Artificial intelligence (AI) and machine learning are pivotal in helping banks and institutions stay ahead of fraud<\/a> and financial crime tactics. However, advanced technologies come with their own set of challenges, especially when it comes to <\/span>model risk governance<\/span><\/a>, a comprehensive and structured approach to managing the risks that arise from the development, deployment, and continuous operation of quantitative AI models.<\/span>[\/vc_custom_heading][vc_column_text uncode_shortcode_id=”146477″]Learn the critical challenges with current AI model risk governance frameworks and how Feedzai is making a difference.<\/span><\/p>\n Many banks face two key challenges regarding AI model risk governance frameworks.<\/span><\/p>\n AI models are not static entities. They self-learn and evolve after exposure to real-world scenarios.\u00a0<\/span><\/p>\n This dynamic nature can be a double-edged sword. On the one hand, it helps catch unexpected anomalies that traditional systems might miss. But on the other hand, it poses a challenge for fraud teams. Banks must ensure that these models continue to produce meaningful results.<\/span><\/p>\n Two primary types of machine learning models come into play here: supervised and unsupervised.\u00a0<\/span><\/p>\n [\/vc_column_text][vc_single_image media=”119107″ media_width_percent=”100″ uncode_shortcode_id=”499131″ media_link=”url:https%3A%2F%2Fhubs.la%2FQ028xt7Q0|target:_blank”][vc_column_text uncode_shortcode_id=”211859″]While the advantage of unsupervised models is clear, it is crucial to maintain vigilant oversight to guarantee their continued efficacy in real-world applications.<\/span><\/p>\n Many jurisdictions, such as the US Office of the Comptroller of the Currency (OCC), mandate the documentation of the entire process involved in creating and maintaining a model that affects individuals\u2019 financial decisions. This documentation is a crucial step in ensuring fairness and accountability in using models. However, there are several challenges to overcome:<\/span><\/p>\n Effective model governance typically requires a dedicated team who is hands-on with the model development and monitoring process. This team should also clearly and explicitly communicate how they use the results.<\/span><\/p>\n This process isn\u2019t necessarily flashy because this governance process doesn\u2019t actively stop criminals. But it\u2019s crucial to always monitor and tune models, as well as demonstrate the validity of the decisions your financial institution produces.\u00a0<\/span><\/p>\n Furthermore, it\u2019s detrimental if it\u2019s done incorrectly. Imagine if your credit bureau cannot prove its methodology for generating your credit score.<\/span><\/p>\n End-to-end documentation of data sources, the intended purpose, development, training, and results of all models is a time-consuming process. A team needs to sit down and type out a multi-page report detailing this process with tables, graphs, and charts to demonstrate the model\u2019s purpose and effectiveness. Some regulatory agencies require this to be done on a semi-annual basis.\u00a0<\/span><\/p>\nThe Challenges with Current AI Model Risk Governance Frameworks<\/span><\/h3>\n
1. Self-Learning and Evolving Models<\/span><\/h4>\n
2. Understanding Supervised and Unsupervised Models<\/span><\/h4>\n
\n
2. Regulatory Expectations for Governance<\/span><\/h4>\n
1. Domain Expertise is Critical (and Time-Consuming)<\/span><\/h5>\n
2. New AI Techniques Carry Risks<\/span><\/h5>\n