Asian institutional investors sceptical about AI’s benefits
“The investment management industry is way behind other sectors’ use of artificial intelligence. Look at auto manufacturers and what they’re doing with self-driving cars." In saying this, Oliver Johnson, Asia Pacific head of investment technology provider SimCorp, has a valid point.
Yet there are a number of valid reasons why institutional investors – and asset managers – are wary of rushing into adopting AI as a portfolio management tool. They centre on reluctance to relinquish human control of the investment process and concerns over creating a less transparent ‘black box’-type setup.
There’s been fast-rising interest in AI in fintech hubs such as Singapore, a technology specialist at a Southeast Asian sovereign wealth fund (SWF) told AsianInvestor. “But I've seen limited momentum on the investment side. I don't know if that’s because we don't know where to focus our energies or if the technology is running ahead of the business being comfortable with its use.”
He has seen “good pick-up in using AI for detecting potentially interesting investment areas,” before humans come in to assess them. “But organisations tend to stop short from letting the computer make investment decisions without there being tight reins.”
The SWF's tech specialist sees even less appetite to use AI for operational processes in investment. “Machine learning and AI can tell you the right answers 98% of the time, but in operations, 98% isn’t good enough,” he said. “Trade confirmation needs to be right 100% of the time.
“So there tends to be less of an appetite to pull the human away from the process on operations, because it’s like a utility: the light has to be on; the water has to run.”
'BLACK BOX' CONCERNS
Key concerns about using AI for investing include its potential lack of transparency and questions about the ability to fix problems if something goes wrong.
“The confidence is just not there yet with AI or machine learning,” said the tech specialist. If a human makes a mistake, they’re accountable for it and can, for instance, fix the formula. “AI is too much of a black box for leaders of an organisation to trust it,” he added. “How do I know you didn't make another mistake in another formula?”
Hence the tech specialist says the sovereign fund he works for is moving cautiously on this front. It's still unproven that AI helps improve returns, he noted. “Some organisations are leaping in a bit harder than others.”
Japan’s Government Pension Investment Fund, Chinese insurance group Ping An and Dutch retirement fund manager APG are among those using AI to enhance their investment capabilities, though not to make investment decisions.
Asset management firms appear to recognise asset owners' concerns. BlackRock's Aladdin portfolio management system does not use AI to make investment decisions for clients, said Akiyoshi Takeuchi, Asia Pacific head of BlackRock Solutions.
“If our clients want to build a proprietary model, they can do that,” Tokyo-based Takeuchi told AsianInvestor. “Aladdin does not make investment decisions, but we open it up to make data more accessible and transparent for clients [who are] using more AI in their investment processes.”
Naturally, regulators are also keeping a close eye on developments. For instance, the International Organization of Securities Commissions (Iosco) is consulting on proposed guidance for regulators in regulating the use of AI and machine learning by market intermediaries and asset managers.
LEADING THE AI CHARGE
Meanwhile, investment technology providers are keen to explore the potential of AI and are pushing ahead with research on it. SimCorp, for instance, has partnered a firm called Alkymi to apply the latter's machine learning technology to complex and unstructured data in the alternative asset space.
“We are beta testing some technology with some clients, where they are ingesting some of these glossy PDF reports using machine learning,” Johnson said. They are using it to save analysts time by interpreting the report and feeding the data directly into the system.
Similarly, some asset managers are at the forefront of using such tech to invest, such as Canada’s McKinley Capital, but even they are keen to stress that computers are not ultimately making the investment decisions.
That is likely to change over time as the technology gets cheaper and comfort levels with it rise, noted the SWF's tech specialist. “I also think there will come a time when the confidence levels are high enough so that more is done using AI for operations.”
“But it's not going to completely replace humans; it's going to be a compliment,” he said. “So there'll be a computer helping me do 90% of the heavy lifting, but there's still going to be the operational person that has institutional knowledge.”