Banks and financial institutions are having to cope with increasingly more onerous fraud threats, according to Hewlett Packard Enterprise’s chief technologist for artificial intelligence (AI), Matt Armstrong-Barnes.
Speaking at Sibos in central London, Armstrong-Barnes said more financial firms were being forced to turn to AI to attempt to stem the rising threats.
“We’re seeing a huge drive in back office automation,” he said. “Fraud detection and prevention is becoming significantly more problematic and AI is one of the only ways we can deal with the tsunami of data that’s hitting us on a daily basis.”
Behind retail, financial services is the second bigger investor in applications of AI, he said.
The list of banks investing in AI capabilities has mushroomed over the past few years. Capital One, Sberbank, NatWest and a range of others have all moved to establish AI-dedicated departments. A report by Accenture published at the start of this year found that banks could see savings between 20 and 25 percent across IT operations, infrastructure, maintenance, and operation costs.
Banks and financial services have been working with complex mathematical sets for a long time, said Armstrong-Barnes, and so AI is a “much more familiar technology for the industry”, at a point in which “we’re seeing a massive explosion in artificial intelligence,” he said.
However, regulators may be quick to move to help shape AI’s development within the industry, which faces hurdles such as a skills shortage as banks and financial institutions look to turn to the technology at speed.
“What if [an AI-ready] chatbot says something to a customer that is outside the context of what you would allow said in one of your call centres?”
“Things like algorithmic accountability and some other disciplines that are developing in the field are really going to help make AI much more widely accepted in the industry,” he said.
Armstrong-Barnes said capabilities are still limited in many cases to narrow applications of AI, such as digital assistants, which can’t apply learning to other applications other than the specific use cases the systems have been built for. However, strong AI – in which a machine can operate like a human across multiple domains – can be expected in the near future. Currently tools at firms’ disposal – including big data, highly sophisticated algorithms, and the ability to choose from a multitude of platforms in order to scale – has driven interest in AI.
“Over the past few years there’s been an enormous adoption of AI and we’re currently finding ourselves at the crest of a wave," he said.