From L to R: Kemp, Fedeli, Laud, and Calay. Source: Morningstar.

Is Artificial Intelligence (AI) worth the money, or is it simply hype when it comes to operations and investment decision-making for asset managers?

These were the questions at the Morningstar Investment Conference on Wednesday in London.

During a panel discussion, Monika Calay, Director of Manager Research, Morningstar, led a conversation on “Reshaping Asset Management” in the industry.

The panel featured Fabiana Fedeli, Chief Investment Officer (CIO) Equities, Multi Asset and Sustainability, M&G Investments; Dan Kemp, Chief Research and Investment Officer, Morningstar; and Sonja Laud, Global CIO, Legal & General Investment Management.

The panel focused on the effects of AI on investment research, operational efficiency, and portfolio performance. One of its main themes of the discussion was the cost-effectiveness of artificial intelligence, both as an operations tool and as one of the main ways of making a better form of investment decision-making.

"The cost of business is going up, so remember

how you do your processes matters."

The rise of AI is relevant to fund operation teams as it could ease the operational admin burden for asset managers by providing additional support for overstretched teams, freeing up staff from mundane tasks. However, it could also come with downsides.

The panel was largely agnostic on AI and its uses so far; yes, it offered exciting new opportunities, but there were still embedding processes to get through before the net benefits were clearer.

“To put it in context, the cost of business is going up, so remember [that] how you do your processes matters,” said Laud. “There are a lot of ways it can help productivity go up.”

However, she stated that the cost element needs to be judged by the output. She posed the interesting question – would adding additional staff members actually be more efficient, and if so, is AI purely being used as a cost-saving measure?

Calay focused the conversation through the wider macroeconomic lens and asked the panel about their thoughts on the wider effect of AI. Recent headlines have highlighted that electricity demand linked to artificial intelligence (AI) is set to double, or even quadruple, by 2030, according to the latest report from the International Energy Agency (IEA). The figures are alarming, particularly for the EU, which has pledged to halve its carbon emissions by then.

In a separate session at the conference, Mark Preskett, Senior Investment Consultant & Portfolio Manager at Morningstar, presented on the macroeconomic outlook for the global economy and focused heavily on tariffs. A key concern he raised around the Trump administration was where the US could build its reshored factories – any AI data centre would require space, electricity and a lot of water. It looked unlikely that it could be in Silicon Valley, where land values are high and the area is famous for its NIMBYism, and situated in California - unlikely to be a first choice for any Trump allies.

This is separate from the environmental cost of AI, which was recently highlighted in the viral story that users of ChatGPT were costing millions of litres of water by saying please and thank you to the chat system. There are multiple concerns about where the water to feed AI will need to come from.

A host of other warnings have also come through in recent months around AI’s use. The UK’s central bank said last month that AI could be a great boon for industries across the economy, particularly finance, but it comes with many risks, some of which are still not fully understood.

In its paper, “Financial Stability in Focus: Artificial intelligence in the financial system,” the Bank of England set out the Financial Policy Committee’s (FPC) view on specific topics related to financial stability. It said that “Operational risks in relation to AI service providers” exist and are “bringing potential impacts on the operational delivery of vital services”

However, AI has also been reported to be a great help for those investing in private markets as it can streamline the large amount of data required in those asset classes.

Productive enough?

Besides the environmental effects, the discussions around AI and its usefulness in the workplace were still very much in the “will they, won’t they” phase for many in the industry.

Laud and Fedeli both said they were using it, but AI’s place in its ecosystem was still fluid.

“AI allows flexibility to learn,” said Fedeli. Yes, she added, it could do non-linear data, but it needs help with unstructured data, i.e., text.

It works well with humans, she said. “Person and machine [together], we’ve learned, is the best combination.”

She also raised the point that no two asset managers are the same; those that have embraced innovation in their internal processes as well as investment decisions will likely have an easier time integrating it than those companies that have been historically slow to adapt to new technologies.

“We have decision making tools in a resource constrained

environment already. It’s called maths."

“The cost of AI will not be homogenous among asset managers,” she said. The more innovative firms will be far quicker, and proprietary data archives, which at the moment, can create headaches for companies not wanting that on the system, could be a boon in the long run and give them an advantage.

Morningstar’s Kemp was less positive in his response and focused on the value-added nature of AI – was it actually delivering on this, or was it a lot of hype?

“We have decision making tools in a resource constrained environment already. It’s called maths,” he said. “What we’re getting AI to do we can already do. It can do research – so can we. It’s good at summarising information – we can do that.”

Despite the passionate opinions, the three figures were largely of the opinion that AI was here to stay and would find its place in the frameworks of the asset management world.

One of the larger changes would be the individual companies, the budget constraints the companies had, their own specific asset classes invested in, as well as the operational structure.

It would be these factors that affected uptake and depth of use of AI rather than a more cohesive approach industry-wide. However, how quickly this personalised approach would occur was up for debate. Many believed that AI was already heavily embedded, but the agnosticism of these senior leaders showed that there was still a lot of work to be done for AI to change the narrative that it was an essential tool rather than an additional resource.

 

Please Sign In or Register to leave a Comment.