A diverse panel from financial institutions, venture capitalists, academia and start-ups considered key questions around AI and financial services, in a session moderated by IBM’s Neil Sahota.
For Stephen Ibaraki, Founding Chair, Global Industry Council & Vice-chair, IP3 Board, International Federation for Information Processing, Canada, AI offers great potential for the streamlining of labour- intensive procedures, such as regulatory compliance scrutinizing or mortgage applications. Procedures that previously took days now take a mere 5 minutes. “Ai is everywhere, embedded, it is a reality of life,” he told delegates.
It’s not just customer facing activities which AI can help streamline; tools such as natural language processing and machine learning also have tremendous potential for enhancing internal processes explained Andy Nam, Chief Information Officer, Standard Chartered Bank Korea Limited, Republic of Korea.
For Woochang Kim, Head, KAIST Center for Wealth Management Technologies/Managing Editor, Quantitative Finance, the value of AI lies with serving “regular people,” whom serving can otherwise be expensive. AI can help “providing the same services to normal people. It’s an unexplored market.” He explained
AI has much to offer in a field such as microfinance, where dealing with the unbanked can be time consuming. “AI can be made more efficient by technology that engages users and connects disparate data points,” explained Chris Czerwonka, Chief Innovation Officer, InvestED, USA. Crucially, AI can also provide local language interaction.
Humans and AI can co-work to achieve effective results, according to Andy Choi of the Republic of Korea’s MoneyBrain, who gave two examples where the AI can receive the message and, if not able to complete the request, humans can finalise it, or alternatively humans can take the call whilst AI analyses it at the same time to help find the optimal solution.
Challenges still abound
No doubt AI offers a great deal of promise but despite a lot of open source tools, more is needed to make it truly accessible, according to Ibaraki. Czerwonka warned of a digital divide: although we are still at a relatively early stage in terms of an AI curve, small enterprises also need access to unlock data sets.
A further challenge for start-ups, explained Czerwonka, is finding the right talent. In today’s lean start-up market, many new entrants might have great visions of implanting a supercharged vision of AI but to do this needs people ready and experienced to build out the technology, and affordably, as well as having some domain expertise too. AI itself may not be that “tough” but the fact that everything is interlinked, such as large data sets or an end to end ecosystem, makes the necessary skillset harder to find according to Nam.
There are also a host of ethical and legal challenges to solve – if AI makes a bad decision, for example, who should be held accountable? As these technologies evolve, so too will the regulatory concerns.
The road ahead
Ensuring people understand AI is crucial in order for it to advance and provide complete value, according to Czerwonka. People also need to understand and be prepared “If the switch goes off and you cannot use AI, you still have the processes and internal controls in place to advance business models.”
For companies starting up in AI, timing is key, but it is, according to Choi a good market to be in if you do it right. A sound strategy, structure, process are essential as a precursor to unleashing AI. AI will not substitute human collaboration, it will be in addition to it. As a precursor to launching AI, explained Czerwonka, it is prudent to make sure sound strategy, process and structure are in place in order to underpin AI. It’s a long journey, not a sprint, concluded Nam. But be sure to check reality, manage expectations and whatever you do, don’t give up!
