Over the past decade, investing through an ESG lens has been catapulted to the forefront of the investment industry. This has led to a boom in new ESG data providers offering different methodologies and views of their optimal approach to assess a company’s ESG/idiosyncratic risk factors. There are thousands of ESG factors/metrics from each of these providers.
At Old Mutual Investment Group, we have conducted extensive quantitative research on ESG data, assessing various ESG data providers. Through this research we have developed a robust understanding of the strengths and weaknesses of the various offerings available.
The explosion in the volume of ESG data has put significant pressure on asset managers to ensure they are using the appropriate data, extracting the optimal value from it and making the best use of these factors across their various investment strategies. Fortunately, with the aid of artificial intelligence (AI), asset managers have the tools to quantitatively and objectively assess this volume of ESG data to find value. AI is particularly good at finding patterns in data, whether it be alpha signals from linear and non-linear combinations of thousands of data inputs or correlations that asset managers need to be aware of when making strategic investment decisions. We have found AI a particularly useful tool to optimise weightings for enhancing our propriety ESG signal.
WHAT ARE YOU TALKING ABOUT?
The application of AI algorithms is constantly evolving. It is now even possible to connect existing live camera feeds, geolocation data, satellite images and sensors to measure ESG factors on a real-time basis!
One area of AI that we expect will relieve significant pressure on asset managers in the processing of the volumes of big data is natural language processing (NLP). NLP allows a fund manager to feed in financial reports and ESG audit reports to get concise summaries. Furthermore, NLP can link to news feeds and warn analysts of topics that are currently “hot” versus topics that do not relate to key subjects. This is especially relevant in tracking ESG events or activities that could potentially impact the profits of a listed company. NLP can even take live online broadcasts in foreign languages, convert them into the language of your choice, rank the tone of the broadcast on sentiment, extract the key themes and summarise the content while also monitoring comments on live feeds from chat groups and social media.
KEEPING IT RESPONSIBLE
It would be naive to believe that AI is the solution to mankind’s ESG problems. Unfortunately, it brings with it ESG risks. This matter is so important that it has a dedicated name in the AI community and is referred to as “responsible artificial intelligence”. Effectively, it helps incorporate the ethical principles set forth by organisations and governments and ensures they are correctly embedded into AI algorithms.
In January 2015, AI experts – including Stephen Hawking, Elon Musk, Google DeepMind’s chief executive Demis Hassabis, Apple co-founder Steve Wozniak and the godfather of AI Geoffrey Hinton, to name a few – signed an open letter calling for research on the societal impacts of AI. The letter affirmed that mankind could reap great potential benefits from artificial intelligence, but pointed out the risks. One such risk is, for instance, autonomous weapons that select and engage targets without human intervention.
Furthermore, to run these artificial intelligence algorithms requires a substantial amount of computing power, which in itself is an environmental risk. The good news is that we could use AI to help minimise its own environmental impact, such as managing data centre cooling, optimising supply chains and minimising ESG risk factors.
NO HUMANS NEEDED
Much of the AI concerns revolve around the social impact to nations. Countries that have a low unemployment rate and struggle to employ people to do mundane and repeatable tasks will benefit from the AI revolution. “Bots” already exist in our daily work and home life, automating tasks in a fraction of the time. The problem is that access to AI is not evenly available. The consequence is that large organisations can use AI equipment to automate the repetitive production of goods without the need for human intervention (e.g. car manufacturing), perform dangerous tasks, such as in mining, or replace people for mundane tasks like crop picking for farmers. In addition, autonomous vehicles will no longer require human drivers and electric cars will reduce the need for filling stations. The question must be asked: What are all these people going to do to make a living in the “new” world?
In light of this, it can be argued that AI could exacerbate inequality within a country and between wealthy and poor nations. Logically, we would need to train up people so that we can replace these “lost” jobs with new ones. This would require more skills using technology, which is again skewed to benefit wealthier and more educated nations.
All of these issues have to be kept in mind when rolling out AI. Implementing responsible AI is by no means an easy task, as it requires balancing the gains in productivity and profits, and the automation of mundane, repetitive and dangerous tasks with the overall societal impact of job losses. One thing is for sure – without the use of AI, the effective management of ESG risks would be practically and operationally impossible. While we can’t afford to be left behind in the AI revolution, as a country with high inequality and unemployment levels, we can implement AI responsibly and be mindful of its ESG implications.