Artificial intelligence (AI) has the potential to improve member outcomes, governance, scheme administration and investment in the pension fund sector, according to a new CFA Institute study, as long as robust safeguards and human accountability remain in place.
The ‘Pensions in the Age of Artificial Intelligence’ report authored by CFA Institute senior affiliate researcher, Genevieve Hayman, suggests that – among other benefits – the burgeoning technology could tailor investment solutions for “individual participants or recommend strategies that would suit each participant based on their personal profile”.
“It may also be easier to incorporate alternative asset classes or expand the range of offerings, allowing participants greater flexibility,” the study says.
But as well as tuning portfolios to member specifications, AI and the associated machine-learning (ML) would probably have further positive impacts on core pension fund investment processes.
“AI and ML could enhance pension portfolio management in numerous ways, including improvements in portfolio construction, optimizing trade executions, and managing portfolio risk,” the paper says.
Indeed, many investment managers are already using, and experimenting, with AI but the CFA report says the now ubiquitous tech will likely spill across all pension fund functions from member onboarding, delivering financial advice to scheme governance and beyond.
“AI technologies may assist multistakeholder interactions and facilitate ease of information sharing, including significant reductions in administrative burden through report summaries, error identification, and prompt responses built on generative AI technology,” the study says. “Increased efficiencies in information sharing and analysis could improve pension boards’ decision making, including investment strategy decisions, and produce timely resolutions to member issues.”
Despite taking a generally upbeat angle on an AI-takeover of many pension fund services, Hayman says, humans would still have some uses: probably.
“Human agents are likely to remain central to pension decision making across the pension value chain. Where humans possess domain expertise or specialized knowledge, those abilities should be enhanced and not replaced by technology,” she says. “AI can function to improve information processing and identification of risks and opportunities. It can also be used to generate efficiencies in the assessment of pension plan status or future outcomes.
“The strong regulatory environment in which pension plans operate, however, necessitates human accountability and strong governance mechanisms.”
Produced under the aegis of the CFA Institute Research and Policy Center, the paper “qualifies for 3.00” professional learning points for those in the CFA community.