Investing in a dynamic world : The impact of AI on the investment landscape

Advances in mechanisation in the late-nineteenth and early-twentieth century automated much of the physical labour performed by humans. Advances in information technology in the mid- to late-twentieth century automated much of the standardised data processing that used to be carried out by humans¹

The next advancement of automation which AI brings is likely to be even more profound. This is because the global ecosystem is in a sweet spot. There are currently 2bn internet users, a forty-fold increase from ten years ago. In another 10 years, that number may rise to 5bn. Today, computer processors are running 40 to 80 times faster than in 2013. The number of connected devices and volume of data are also on an exponential path. By 2020, there could be as many as 50bn connected devices. By 2025, 1tr gigabytes of data would have been generated, ten times more than what was generated in 2016. Data is the “oxygen” of AI as it enables machine learning, one of AI’s core pillars, to be effective. All this helps to create the infrastructure that AI needs to deliver real value to serious adopters and be a powerful force of disruption.

AI is expected to add USD13tr to global Gross Domestic Product (GDP) by 2030, boosting global growth by 1.2%². The impact will come in the latter part of the period and will be uneven depending on the adoption rates in the different countries. Within Emerging economies, the impact will vary greatly subject to the stage of development of their AI ecosystems. China will do well given its AI focus as evident from its “Made in China 2025” strategic plan. China aims to become a world leader in AI by 2030³.

AI will also have very meaningful and differentiated impact across industries. A study of 400 case studies by McKinsey shows that 67% of the time, AI can add incremental value over traditional analytical techniques⁴. In revenue terms, this represents USD3.5tr to USD5.8tr of additional annual value or between 1 to 9% of 2016’s revenue. See Figure 1.

Fig 1: AI Impact⁵

AI’s impact is higher for industries that generate a significant amount, variety and velocity of data, and where traditional analytical techniques are already being used such as the retail and travel sectors. The application of AI in these instances can generate additional insights and lift productivity.

In the healthcare industry, AI has revolutionized medical diagnosis and introduced a new world of virtual nurses and assistant surgeons. In a 30-minute test, Alibaba’s Doctor You programme achieved a more than 90% accuracy rate in detecting lung sarcoidosis. It took four doctors two to three hours to process the same amount of patient data. Using AI for medical diagnosis can help alleviate the workload of doctors in China, where there is an acute shortage. In the US, Accenture estimates that AI applications can create USD150bn in annual savings for the healthcare sector by 2026. AI will also allow doctors to improve their time management. It is estimated that over 40% of physicians’ time is spent on administrative tasks. AI will help free up some of that time and allow doctors to focus on higher value-added areas such as patient management and surgery. AI’s impact on drug design and testing will be revolutionary, amplified by the usage of nanotechnology.

In the logistics and transportation industry, it is estimated that Machine Learning (ML) and AI can potentially disrupt USD500bn of logistic expenses annually⁶. AI can optimize routes by incorporating traffic updates, road closures and weather forecasts. This improves fuel efficiency and reduces delivery times. By using AI to predict congestion and weather-related problems, airlines can avoid costly cancellations. Algorithms are also able to assist warehouse workers with order picking, hence lifting productivity.

Other industries including banking, retail, telecommunications, travel and insurance are expected to enjoy significant benefits from adopting AI applications. Substantial value opportunities for AI exist in the marketing and sales functions of these industries where personalised product recommendations can be delivered by analysing customer data. By incorporating customer touchpoints and content attributes, AI can predict customer demand and trends.

Besides lifting productivity and growth, AI will help to lower costs. Its ability to reduce costs in the logistics industry while raising productivity has resulted in companies producing goods and services more efficiently and cheaply. This is evident in the retail industry where AI’s predictive power shortens the feedback loop and results in better warehouse and inventory management.

In other industries such as the Telecoms sector, AI-powered chatbots are replacing human customer service representatives. With the cost of a chatbot query costing 1/8th that of human interactions, the cost savings are tremendous, particularly since personnel costs make up 40% of telecom operators’ total operating expenses. For traditional lenders, AI and ML can meaningfully reduce credit risks by identifying accounts at risk and executing credit line reductions/eliminations to avoid charge-offs and loan losses. Estimated cost savings in this area can be as much as USD13bn per year by 2025⁷. There are also multiple avenues for oil & gas companies to leverage AI. AI applications can better target oil and gas reserves, thereby lowering oil field development costs. Preventive maintenance improves equipment uptime and lowers production cost. Lowering the oil and gas industry’s capital expenditures, operating expenses and inventory costs can potentially result in savings of about USD140bn over a 10-year period⁸.

By adopting digital technologies, it is estimated that European telecom operators, for example, will be able to boost Earnings before Interest, Taxes, Depreciation and Amortisation (EBITDA) margins by 4% and EBITDA growth by 2% annually over the next 10 years⁹. Asset managers, meanwhile, may see a return of 2007’s peak margins (41%) by using ML and AL in data analysis, identifying investment opportunities and trade executions¹⁰.

It is, however, too simplistic to expect that the advancement of AI will only bring benefits.

Besides altering geopolitics in a potentially significant way AI’s “dark side” must also be considered in the context of the impact it will have on jobs and how it will dictate the skills needed for tomorrow’s workforce.

The Organisation for Economic Cooperation and Development (OECD) estimated that 14% of jobs across 32 OECD countries are highly automatable and another 32% have a significant risk of automation. Jobs that involve repetitive tasks are most at risk and such jobs that remain will likely experience intense downward wage pressure. While some of these job losses will be offset by new jobs, the disruptions in the workforce must be taken seriously. Ultimately while AI presents many positives for a significant part of the global population, it will also create uncertainty and instill a sense of insecurity for many. It is paramount that political leaders and policy makers acknowledge and address these issues.

Education holds the key in ensuring that workers are equipped with the right skill sets to thrive in the new era. Productivity improvement contributed to more than 50% of US economic growth in the last 40 years, with education accounting for 30% of the productivity increase¹¹. Governments must make long-term investments in education and provide continuing education to help mid-career workers. Selected US universities have started to combine coding with other subject majors. The Chinese Ministry of Education, earlier this year, issued a detailed action plan on how China will promote AI development in universities. This includes strengthening the knowledge of basic AI theories and promoting greater interdisciplinary efforts between AI and other subjects¹².

As the development of AI and its application advances rapidly, the interaction between humans and machines will increase dramatically. Understanding the algorithms that guide some of this machine development will be crucial as well as having a multi-disciplinary approach to working and learning. It is also imperative to think about the skill sets needed to win in the new AI-enabled economy. In his book “Humility is the new smart” Edward D. Hess stresses that to stay relevant, humans will need to play a different game. They will have to be critical, creative and innovative thinkers as well as engage effectively with each other.

Remembering Bill Gates’s quote: “We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten”, we should expect the impact of AI on the competitive landscape to increase dramatically over the next decade. Traditional views of winners and losers at the company, sector, country and geopolitical level will be challenged as will the divide between emerging and developed economies, depending on each’s ability or scope to adopt AI.

Despite its dark side, we believe that AI will be positive for capital markets. Cost deflation, including a lower cost of capital, as well as the lift in productivity and growth should impact financial markets positively and meaningfully. Investors, however, will have to accept new sources of risks and volatility, as the drivers of competitiveness and what we perceive as relevant skill sets shift rapidly.

In the short term, markets are likely to be choppy amid interest rate normalisation, heightened political volatility and shifting growth dynamics. This environment provides investors with opportunities to reposition their portfolios. Patient investors who accumulate exposures in companies that embrace the advantages which AI brings, will benefit.