Talk
Buying AI Doesn’t Mean Outsourcing the Risk
What African banks, executives, and regulators must know before signing
Many banks, fintechs and regulators in Africa acquire AI rather than build it. The decision then moves from the engineering team to the board, the risk function and the supervisor, and it is taken in a contract.
Buying well takes capacity too: knowing what data the model was trained on, whether its decisions can be explained, where the customer’s data goes, what the system costs over time, and what the institution keeps when the contract ends.
The premise is simple: outsourcing the technology does not outsource the accountability. A credit decision, a fraud alert or a compliance flag produced by a vendor’s model is still the institution’s decision, and the institution answers for it.
It closes on the questions to put to a vendor before signing, and on what supervisors can reasonably ask of institutions that buy AI. The contract is a governance document.
Questions this talk answers
- 01What changes when a bank, fintech or regulator buys AI instead of building it?
- 02What should a buyer ask before signing: training data, explainability, where customer data goes, cost over time, what it keeps at the end?
- 03Who answers for a decision made by a vendor’s model, and what can supervisors ask of institutions that buy AI?
Want this talk at your event?
Send your date, your audience and the format you have in mind. The talk will be adjusted to your room.
Invite Hamid with this talk