Westpac is turning to a finance-specific large language model (LLM) created by Kasisto for its generative AI experiments, rather than using ChatGPT, which the bank’s CTO, David Walker, has described as “overhyped.” Walker’s latest comments suggest that ChatGPT did not pass Westpac’s internal risk assessments. The LLM, which is specifically trained for financial services, is seen as “more accurate, safer, and intuitive for our customers and employees, and still delivers a ChatGPT-like experience,” according to Walker. Westpac’s security teams are monitoring developments in generative AI to assess how they could affect the bank’s risk posture.
Westpac seeks generative AI experiments outside of ChatGPT
