Every new technology discovered does one of four things: it enhances, augments, replaces or eliminates. Innovation is a complicated beast – “creatively destructive”, as the economist Joseph Schumpeter said. It’s typical that, just when we think we’re comfortable, a new way of doing things throws up unexpected consequences that can disrupt, or even dismantle, entire industries.
AI can be considered a ‘general purpose technology’. It has the potential to drastically reshape the entire fabric of the economy, and that includes asset managers. The potential for large language models (LLMs) to integrate with and complement operations, but also play a role in primary functions of asset management, such as portfolio construction, due diligence, research and client journeys, presents a new challenge in a challenging macroeconomic environment. As difficult as it is in a time of significant cost pressure, critical mass will only be sustained through prudent internal (as well as external) investment, which lifts the industry’s productivity potential in the long-run.
Adapting to AI at a time when we are not certain of its full extent of application, is likely to result in unevenly distributed benefits. Early adoption often creates a mix of new risks – especially if we are overly dependent on or allured by a new technology. AI hallucinations, where misleading results based on insufficient data can cause massive damage if left unchecked, is one example and there are others. Cybersecurity risks, caused by deep-fakes and AI attacks on IT infrastructure, are particularly concerning for the financial sector.
But we know that, by-in-large, AI will have an expansionary effect on productivity, both in terms of the potential quantity and quality of the advice we can deliver.
We know this, because there are already cases of where natural language generation (NGL) is automating cornerstone processes, such as portfolio comments and multi-factor portfolio allocation. Within that, there are repetitive tasks like retrieving documents for financial analysts, translation and summary generation from structured and unstructured data that are already part of AI’s domain.
Data quality controls to ensure the consistency and reliability of ESG data, sentiment analysis on stocks – there are demonstrable proof points across reporting, operations, and research where AI is widening the productivity frontier of asset management.
Speed, especially the pace at which data is synthesised and used to inform investment content, requires strong investments and scale. Access to information has always been part of the moat, but the ability to deliver more comprehensive investment appraisals alongside faster and more educational explanations on portfolio performances with a wider field of considerations included, is set to elevate the overall client experience.
As an industry we should also anticipate that AI’s ability to create highly customised products will change customer expectations. We’re already seeing this in the context of non-financial objectives and their inclusion in asset allocation, where automated allocations can provide well-tailored solutions addressing investors’ diversification needs on the basis of sustainability.
Global society as a whole has wider issues to contend with when it comes to living with, and embracing, AI. At least in the early phases of its commercialisation and diffusion into labour markets, we may well find ourselves in a position where technological upheaval contributes to stagnating wages at the same time as an uplift in economic output. Job displacement and creation, a widening gap between developed and emerging markets, and prospective policy intervention, are all contingencies at this time and AI regulation must quickly establish its boundaries.
But for asset management and other financial services that are already facing growing demand for efficiency, there is an urgent need to embrace the alien. For like it or not, we cannot see a future without it. It’s here now. And there is no way to return to the world before.
By Edouard Legrand, chief digital officer, BNP Paribas Asset Management
