The proper use of Artificial Intelligence requires Compliance Intelligence. You as a compliance officer need to understand what your bank is doing and how. Part of this is vendor due diligence. When marketing wants to use a vendor to advertise on social media, it’s important for everyone to understand where the criteria comes from on the selected target market. There could be any one or more prohibited bases used in the marketing program and one bad apple could spoil the whole bunch. 

- If AI algorithms are trained on biased data, they may inadvertently perpetuate or exacerbate existing biases. This can result in discriminatory lending practices, where certain individuals or groups are unfairly treated when applying for loans.

• While AI can enhance fraud detection capabilities, it could be misused if the algorithms are not appropriately trained or updated. Failure to detect fraudulent activities may result in financial losses and could violate banking regulations related to security and fraud prevention.
• If AI is used in debt collection processes, there's a risk of implementing overly aggressive or unfair practices. Automated communication that violates consumer protection laws, such as constant and unsolicited contact, could lead to legal issues.
• AI can assist in identifying patterns indicative of money laundering, but improper implementation or lack of oversight may result in the failure to comply with AML regulations. This could lead to legal consequences for the bank.
• Some AI systems, particularly complex machine learning models, lack transparency and are difficult to explain. If a bank cannot provide clear explanations for the decisions made by AI systems, it may violate regulations that require transparency in financial transactions.

The above bullets are all valid. I asked ChatGPT how a community bank may mis-use AI, and it told on itself. But sometimes it is wrong and you need to recognize when it is.

This is the beginning of a new era and AI will be improving banks bottom lines. This is not far from the days when loan contracts were fill in the blank forms and the Fed box data came from TILA manuals with table after table of disclosures, suddenly being replaced by computers that were very precise. 

But we need to understand and be able to explain and yes, justify, the data used and our actions taken. This is your opportunity to begin your journey and learn some fundamental of controlling your AI risks.

WHO SHOULD ATTEND: While targeted at compliance topics, compliance is everyone’s job and this program will discuss fundamentals applicable bank wide as AI technology may be employed and risk is managed. Compliance officers, auditors, management, marketing and anyone involved in vendor due diligence should attend.