AI can process more information than ever. Judgment still decides what matters.

Can AI be trusted with investment research?

Artificial intelligence in asset management is moving from experimentation into everyday research.

Many investment firms are no longer asking whether to use AI. They are deciding where it belongs in research processes, how its outputs should be governed and where human judgment remains essential.

Yet one theme emerged repeatedly during Bloomberg’s Investment Management Summit in Singapore: investors are becoming more comfortable using AI to support research, but responsibility for investment decisions still rests with people.

Research teams are asking different questions

Most firms began their AI adoption with straightforward tasks such as summarizing earnings calls, research reports and market news. Now the discussion is shifting toward idea generation.

Participants highlighted AI’s ability to identify patterns across large volumes of information. Rather than simply reducing reading time, AI can help analysts connect developments across industries, markets and companies that might otherwise be missed.

As Bloomberg AI specialist Jun Yang Teh explained, “AI’s real value is in identifying patterns at scale across multiple information sources.” But he also warned that AI can sometimes “connect things that aren’t really there,” making analyst oversight essential.

That tension is shaping how investment teams use AI: to surface ideas and accelerate analysis while leaving analysts to assess their significance.

AI can identify patterns. Analysts assess their significance.

For all the attention on AI, investment professionals remain clear-eyed about its limitations.

Jason Low, a former investor and now Asia industry specialist at Bloomberg Intelligence, challenged the notion that analysts are being replaced.

“There’s a lot of talk about AI replacing humans, but honestly it hasn’t happened,” he said. “AI usually captures two or three of the key points we need in a piece of research, but we still need the human to complete the picture.”

Markets are influenced by historical patterns, but investment decisions are often made during periods with limited precedent. Corporate restructurings, regulatory changes, geopolitical events and technological shifts rarely follow a script.

AI is effective at recognizing patterns. Humans remain responsible for assessing relevance, weighing competing evidence and determining what matters most.

Trust matters more than speed

One of the more revealing discussions at the summit centered on what investors value most when evaluating AI tools. Trust, rather than speed, emerged as a priority.

For institutional investors, particularly large asset owners, governance, security and transparency continue to shape adoption decisions. Research teams need to understand where information comes from, how outputs are generated and whether conclusions can be verified.

That requirement is especially important in investment management, where accountability cannot be delegated to an algorithm.

The goal is not unquestioning acceptance of AI-generated outputs. It is creating a process where AI can challenge assumptions while preserving human accountability.

Research is only valuable if people can use it

As AI adoption expands, another issue is attracting attention. The challenge is ensuring that research can be found, shared and acted upon across an organization.

Many firms still operate with knowledge spread across emails, spreadsheets, notes and disconnected systems. Valuable insights are produced every day but are not always accessible when they are needed most.

AI can help teams search and connect information across those sources. By connecting research, market data and institutional knowledge, firms can make information easier to discover and use. Instead of spending time locating information, portfolio managers can spend more time evaluating ideas and testing assumptions.

The role of the analyst is changing, not disappearing

Perhaps the most notable takeaway from the summit was how practical the conversation around AI has become.

Few participants argued that AI should take responsibility for investment decisions. In fact, when attendees were asked where AI should never operate independently, the final investment decision emerged as the clearest boundary.

That boundary points to a broader shift in investment research. As information becomes easier to access and analysis more automated, the analyst’s role may shift further toward interpreting evidence, testing assumptions and exercising judgment.

AI can change how research is conducted. Responsibility for the investment decision remains with people.