Commonwealth Bank Carefully Addresses ChatGPT Bias – Finance – Software

Commonwealth Bank actively evaluates the risks of experimenting with or using conversational AI platforms such as ChatGPT, including how reliable their output is.

At a Trans-Tasman Business Circle event, Chief Data and Analytics Officer Andrew McMullan shared the challenge of understanding what generative AI is using to generate answers to questions he (and banks Overall) said they were very “aware”.

“We need to be aware of what the AI ​​is learning and what data it uses to come up with answers,” McMullan said.

“Everything ever digitized and every ‘ChatGPT’ is potentially biased data in terms of when it was written and learning about the story at that time.

“From a CBA organizational perspective, we are focused on fully understanding the use case. We have our own policies [for experimentation and use].”

McMullan said one of his roles within the CBA is to act as an AI risk steward and ensure that “principles and policies and all use cases for AI are thoroughly considered before doing anything with AI. It is to be done,” he said.

“I have always believed that the use of AI and advancing science and technology expresses an organization’s culture.

Earlier this month, CBA told iTnews The AI ​​Labs research team said,Actively participate in recent advances in generative AI, ChatGPT and similar technologies are emerging. “

The bank is exploring use cases that can enhance its customer service delivery. customer engagement engine (CEE).

“For us, the use cases we have are very focused on our purpose of trying to help Australia and Australians have a brighter future and the areas we really focus on. is guessing. [are] Service,” McMullan said.

“How easy can it be for frontline staff to serve customers and get answers to customer questions every time they call a call center or visit a branch?”

Beyond customer service, McMullan said AI has also proven useful in monitoring IT systems and reducing data duplication.

He also said banks continue to invest in academia and their skills working with AI and data.

“We are seeing a shift in the skills needed to be able to provide great service to our customers every day, so we are training people who used to work on the front lines to do engineering with us, or do analytics. Or you can even get into machine learning and AI if you have the desire to re-build your skills in the machine learning and AI space,” McMullan said. increase.

“I’m excited about what we can offer over the next decade with AI, an advancement in machine learning,” he said, adding, “As society becomes more digital, our customers’ experience in Australia will continue to grow.” I think there will be better features available to support the whole.”