How ASB is using AI to beat AI

As scammers and bad actors become even more sophisticated in using artificial intelligence, banks need to meet that challenge and the needs of customers, says ASB chief executive Vittoria Shortt. 

ASB is using AI in a variety of ways, helping to streamline and speed up business processes and enhance the experience it offers customers. Using systems embedded with AI to identify potential fraud is another area where big wins are made.  

“Scammers are constantly evolving their approach and using technologies that are making scams harder to detect. We too must make sure we are using technology to combat them,” says Shortt.

“AI has been successful in detecting and preventing scams. We already use a lot of AI in our fraud systems, such as being able to auto block fraud payments, but we don’t have the ability to prevent or block everything.”

A lot of these scams target average New Zealanders who are duped into parting with their money by unscrupulous actors, using increasingly convincing fake images, video or audio recordings. 

“I’m very worried about scams, in particular,” Shortt says. “When they first appeared on our radar, some were easier to spot – thanks to spelling mistakes and poor graphics. Now they are becoming so sophisticated, along with the way scammers target people. 

“There is no silver bullet, so we have got to make sure we are continuing to up our game in a multitude of ways – from stopping customers receiving scam messages, via text, email or other channels, to making them more aware of scammers’ techniques – such as the use of deepfakes – and reminding them of steps to take to protect themselves.

“That’s the awful part about AI. It is breaking down trust in society. We have got to make sure we are continually investing because nothing is standing still.”

There have also been a lot of claims about AI and the impact it can have on productivity, with estimates of between 20 to 30 per cent improvement across sectors – as a starting point.

Shortt has seen this in action, particularly at contact centres where customers “can ask anything and everything”, and it can be difficult for a person to have all the answers to hand quickly. AI can quickly produce policies, relevant content and material in a consistent way. 

“I’m a big believer in AI and Gen AI. I see an opportunity for it to help us to be more efficient, and to communicate and digest information more effectively. But safety has to come first. 

“This is where board governance comes in. It has got to start with policy. We have looked at some globally recognised principles around AI and are using those in our policy,” she says, listing key points as:

1. Human, social and environmental wellbeing: AI has to benefit people, and you have to know how and articulate that.

2. Fairness and transparency: You need to make sure there is no discrimination or bias and, if there is, how you are going to manage it. You must be able to understand and explain the outputs. 

3. Privacy and security: That AI is easy to understand, being used in your data set and doesn’t go externally.

4. Reliability and safety: Large language learning models must be reliable throughout the lifetime of their use. You can’t just launch something, let it loose and then sit back.

5. Accountability: You need to know there is a person in the team who is accountable for every AI model. Accountability is a critical part of any policy framework.

Shortt says the bank has had a lot of discussions to ensure the right foundations are in place. “You can’t just assume everyone knows everything about AI. For the Board, it was about starting with the foundational elements and then building their knowledge base. 

“For example, we took the Board through how the audit team is using Gen AI, including the tools, so they could see and understand individual use cases and how it is being applied in practice,” she says, adding that external parties have also expanded their thinking and ideas. 

Referencing Jeff Bezos’ advice that “you can build a business strategy around the things that are stable in time”, Shortt looks at AI and thinks about what is unlikely to change while everything else is changing. 

“The other thing that is always on my mind is humans also need humans. So, it is about finding the good in all of this.”