A sustainability lens on AI

Although still in its infancy, Artificial Intelligence, better known as AI, is already transforming industries.

A sustainability lens can help to understand the potential benefits and pitfalls of that change, says Royal London Asset Management fund manager George Crowdy.

Source: https://commonslibrary.parliament.uk/research-briefings/sn02783/ 

Accelerating the shift to the cloud

Stock markets today are rewarding AI enablers such as semiconductor companies and parts of their supply chain. At the same time, because nearly all AI applications will be done in the Cloud, it will accelerate the shift towards cloud computing, Crowdy said. “According to Amazon, about 85% of computing workloads are still on-premise, so this shift has much further to go and the big cloud players like Microsoft Azure, Amazon’s AWS and Google Cloud will be beneficiaries.”

Ethical implications of AI

But there are growing concerns about some of the ethical implications of generative AI in particular. “This is different to previous technological shifts because humans will have much less of a role in the output that it produces. That comes with risks, and our role as a sustainable investor is to ensure that companies are innovating in a responsible way.”

This means ensuring that model and product development processes have strong governance from a broad selection of people with different backgrounds, to guard against inherent biases in the way generative AI produces responses and outcomes.

Because we are in such an early stage of AI growth, a lot of responsibility is being placed in the hands of a relatively small number of companies that have the capability to develop products at scale.

“Even before generative AI, there was a real challenge determining what was genuine and what was fake content. Generative AI will clearly make it even more difficult for end users to know what is genuine and what isn’t.

“It's a really fast-moving space and regulation will struggle to keep up with these new technological shifts. This is where, as sustainable investors, we want to see companies setting standards and guardrails to ensure that outcomes are as responsible as possible.”

Adobe: responsible data sourcing gives a competitive edge

One example of this is Adobe working on text-to-image, and text-to-video, generation. “There is a huge risk of inherent bias in this area: if you prompt AI to show you a picture of a doctor, you don't necessarily want to be presented with a picture of a white British male.”

The way Adobe has handled this is to use its own library of stock images to train its AI – unlike other text-to-image models, which are often simply ‘scraping’ the internet for content. This gives Adobe a genuine competitive advantage because its customers know the models are being developed responsibly, Crowdy notes.

Using resources wisely

When it comes to enablers like Nvidia, one of the key sustainability issues is around the resources needed to produce computing advancements. “Making chips is hugely resource-intensive from an energy and water perspective. The vast majority of advanced semiconductor chips are produced in Taiwan, which is a water-constrained region.

“Here we want to see companies innovating to reduce the resource dependency of their products. We want to see a company’s next product range using fewer resources, but doing more. Nvidia’s latest semiconductor chip provides up to a 30X performance increase and reduces cost and energy consumption by up to 25X versus its predecessor for inference workloads.”

Data source: https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-economic-potential-of-generative-ai-the-next-productivity-frontier#key-insights

Managing operations

While much of the AI story so far has been focused on products, an even bigger long-term opportunity could stem from companies using AI to manage their operations, Crowdy believes.

“It's very early in the process but already we are seeing automated customer-service centres rolled out in almost every industry,” Crowdy said.

“Any industry that has manual repetitive tasks that involve data – for example, the banking industry – could be huge beneficiaries. If you're not doing smart things with data when it comes to assessing people for credit in the banking industry, you will quickly be left behind.”

He gives the example of Indonesia’s Bank Rakyat, whose specialisation is in micro-lending, with an average loan size of around $4,000. “They are using AI to better understand people’s credit history and how things like weather conditions might impact a customer’s ability to repay loans.”

Physical infrastructures

Other ways to play the AI theme in a sustainable way include physical infrastructures. Hundreds of billions will be spent on building data centres over the coming years and making these environmentally friendly will be hugely important.

“One company we’re investing in produces the world’s most environmentally friendly and advanced heating and ventilation air-conditioning (HVAC) systems. Cooling in data centres is vitally important. One of these systems can cut energy bills and cool a building more efficiently by using more environmentally friendly refrigerants so it’s a win for everyone.”

​

The value of investments and the income from them may go down as well as up and is not guaranteed. Investors may not get back the amount invested. This is a financial promotion and is not investment advice. The views expressed are those of the author at the date of publication unless otherwise indicated, which are subject to change, and is not investment advice.