Look at collective intelligence to deliver impact

In complex decision-making environments, where uncertainty about the future is significant, technology on its own is not yet ready to supplant humans. In 1859, Charles Darwin observed that species survive change through adaptation. While we can agree it has become rather simple for Google AlphaGo to beat a go master, the odds are very different when it comes to predicting what something as irrational as financial markets will do over the next year. The disappointing performance of many flagships quantitative funds in 2020 is a good illustration of the current limits of artificial intelligence (AI). In contrast, numerous discretionary hedge fund managers adapted well to the formidable challenge of the pandemic.

Collective intelligence can offer solutions. This is not a new concept: in 1906, the British statistician Francis Galton — a cousin of Mr Darwin — pioneered the concept of “the wisdom of the crowd”, identifying that the aggregation of judgement in large groups would provide better forecasts than the best predictors in that group.  

Not all crowds are wise, however. For the aggregate result to be greater than the sum of its parts, and for collective intelligence to emerge, several key features are needed: a defined group; the presence of independent thinkers; a diversity of viewpoints and perspectives; and a structure where decisions are not centralised.

In a world of big data, a common error is to focus on group size when it comes to extracting intelligence – the premise being the larger the group, the more information, the better the decisions. However, while size is power, with larger sizes come increased noise. Instead, it is the organisation of the group that makes a critical difference. Independent thinking, or individual expertise, should not be overlooked. Diversity of opinion and size eventually compensate for lower expertise, but the more diverse expertise a group can access, the smarter it will get.

“The essence of intelligence is the ability to predict,” said Yann LeCun, founder of Deep Learning and Facebook’s chief AI scientist. Organised well, a group of humans can be remarkably good at prediction. 

We have witnessed this first-hand at the company I founded seven years ago, 1859 CI — a cloud-based platform that links up professional investors. The aim is not to forecast where the market will be, but to identify investment opportunities that will most likely outperform in the future. Blending human qualities with technology, it is possible to build models that can produce good risk-adjusted returns to investors while also generating a regular source of funding for, say, impact initiatives, which go towards meeting the UN sustainable development goals and where the ability to deliver short-term results may be less obvious.

The big-data trap can be avoided by focusing on capturing a bespoke dataset in a selected community of experienced buy-side professionals across a range of trading styles, asset classes and regional locations. Reliable, detailed knowledge of behaviours and qualities of this community leads to better judgement. This judgement, otherwise, would be just an average of a small sample of opinions. Again, strength is in pattern details, not size or majority. Vetting which professionals form the community helps ensure the presence of independent thinking, as well as diversity.

Driving capital towards investments that have longer-term benefits can be challenging if results are less obvious in the short term, but it is not impossible. Combining technology and the understanding of human behaviour will help the investment community’s contribution to the broader economy. 

Patrick Bauné is the founder and chief executive of 1859 Cloud.