Institutions embrace AI for portfolio optimisation potential
Having recovered from the initial existential dread brought on by the advent of generative artificial intelligence (AI), investors are now learning how the technology can enhance their work.
When the technology firm OpenAI introduced ChatGPT in November last year, there were many dire predictions about the dangers it would pose, potentially making humans obsolete and perhaps even extinct.
Goldman Sachs estimated that 300 million jobs could be lost over the next 10 years, but the bank also predicted that the combination of significant labour cost savings and new job creation raises the possibility of a productivity boom.
GOING MAINSTREAM
One year on, and investors are already exploring these new efficiencies, with OpenAI suggesting that as many as 80% of staff at large US companies are using ChatGPT in their work.
In the financial sector, asset managers and their pension fund clients have been turning their minds to how AI can benefit their investment teams and scheme members respectively. Some are actively encouraging staff to use AI.
For example, at Vision Super in Melbourne, chief investment officer Michael Wyrsch told AsianInvestor that the fund has formed an AI working group to establish guidelines for its use.
“We believe that AI has exciting possibilities. We are encouraging staff to innovate and play with tools like ChatGPT, while restricting any use of proprietary information or member data to safeguard privacy.”
Unisuper CIO John Pearce sees potential for AI to provide analysis for the optimisation of portfolio and investment risk management. He is also looking at the technology as a means to deliver better service to scheme members.
“We have people across the organisation, including the investment team, working together to determine how the latest tools can be best put to use, safely, and while withholding our strict standards of data security and privacy.”
Meanwhile, Bloomberg users are now able to access an AI platform capable of making sentiment analysis, news classification, and other financial tasks.
Pearce said that AI is already helping to improve their investment research. “AI has assisted with searching, summarising, explaining technical information, and translation.”
“Any customer-facing industry is going to benefit from it, and any industry that requires research is going to benefit tremendously,” Hong Kong entrepreneur and family investor Timothy Tsui said, describing his AI experience thus far.
“Utilising AI to summarise a dense legal document or a white paper is a time-saver at the very least. Do I have to read 40 pages of risk factors in an IPO prospectus, or can I get a summary of the main risks?”
The AI can also handle more complex customer enquiries, allowing a shift of human effort towards experimentation and innovation, said Tsui.
The business transformation could potentially be so dramatic that the investment case is undeniable, he added. He used the example of how AI is helping mining companies map for minerals and improve their new mining site search capabilities.
“What a lot of people don’t realise is that if AI improves your success rate in R&D or in solving customer-related issues from 5% to 20%, people will be willing to pay a lot of money for that.”
BUSINESS RISK
On the question of data security and reliability, law firm Linklaters has taken a clinical look at some of the potential pitfalls for organisations exploring AI.
“AI is a constantly evolving disruptive technology posing novel ethical and practical challenges, and its deployment in financial services raises complex legal and regulatory issues,” said the firm in its latest newsletter.
It notes that there are AI-specific risks such as employee experimentation, unreliable outputs, and limitations of knowledge, which may be amplified by the use of generative AI and could lead to financial loss, reputational damage, and legal sanctions.
Alternative asset managers and any investors using more advanced AI techniques may be exposed to the risk of outputs that may not be explicable as a function of their inputs, said Linklaters.
Some experts have identified a trade-off between efficacy and explainability, the firm said.
"Processes of reverse engineering can sometimes be used to draw conclusions about the properties of so-called ‘black-box’ algorithms, although these will not provide complete transparency. Firms deploying AI solutions will need to consider these novel features in determining the adequacy of their existing governance, oversight, and risk-management frameworks,” in continued.
Overall, the views being expressed now reflect a more sober assessment of the threats and opportunities posed by AI.
Speaking at the recent AsianInvestor Insurance Asset Management event in Hong Kong, Courtney Wei, deputy general manager at China Life, said, "We have tended to overstate the direct effect but understate the long term potential."