The AI evolution: Where are we heading?

Artificial Intelligence (AI) has dominated headlines this year and sparked a rally in technology stocks. Jeffrey Lin, Head of Thematic Investments, believes that generative AI represents a significant technological advance, with potential benefits extending across a broad range of industries. So, what is the exciting spectrum of investment opportunities that could be created by this disruptive new technology?

2023 was arguably the year that AI went mainstream. Over the past few years, AI has become increasingly prevalent in everyday life, used in voice assistants such as Amazon’s Alexa and Apple’s Siri; customer service chatbots; personalised online recommendations for shopping or viewing choices, and much more. However, this year, there has been a dramatic surge in interest in AI and its capabilities thanks largely to ChatGPT.

OpenAI’s large-language model (LLM) chatbot, ChatGPT, was made freely available in November 2022 and became one of the fastest-growing applications ever, reaching 100 million users in just two months^[1]. LLMs are advanced AI programmes that are trained on large amounts of text data to understand human language and produce meaningful, natural responses. The technology is designed to make communication between humans and computers more like human-to-human conversations.

ChatGPT sparked a wave of excitement around AI because it enabled anyone with access to the internet to experience generative AI’s enormous potential. In its response to simple prompts and questions, it has been a powerful demonstration of how generative AI can enable computers to ‘think more like humans’ and create original text, images, video, music and code.

“Generative AI can enable computers to ‘think more like humans’ and create original text, images, video, music and code.”

Tech companies have been working on AI projects for years, but OpenAI opened the floodgates with the release of ChatGPT. It has been swiftly followed by a host of other open-source generative AI tools, including Google’s Bard, Meta’s Llama and Baidu’s Ernie Bot, as large tech companies seek to capture a core share of the market.

The hunger for data

LLMs rely on vast amounts of data to be able to produce new content, which is leading to growing competition for datasets. Data has already been hailed as the new oil. Generative AI might as well be the internal combustion engine of our age that helps unleash its value.

Businesses that own proprietary datasets will become increasingly valuable, especially since those datasets are required to create the generative models. For example, Adobe, a creative software company, has used its database of hundreds of millions of photos to create Firefly, a product that generates AI-powered content with a simple prompt. Sources of news and images are also being approached by AI companies to incorporate into their models.

In the future, we are likely to see businesses focus on training models on their own proprietary data to create bespoke AI applications. Given that many firms’ datasets are located across different systems and may not necessarily be in the best format, this could require the help of IT services companies, who can advise firms how they can implement AI effectively and achieve efficiencies in their operations and ensure data security.

A disruptive technology

We believe AI is now entering the next phase of its evolution driven by generative AI and LLMs. In our view, the capabilities and applications that AI now offers has huge potential to transform businesses, disrupt entire industries and even alter human lives.

Using computers to enhance human decision-making and perform repetitive tasks could help businesses improve efficiency and productivity. Research by consultants McKinsey estimates productivity improvements from generative AI and automation could add between US$2.6 trillion – US$4.4 trillion annually to the global economy^[2].

And the benefits of generative AI are not confined to the tech industry. We believe a broad range of industries have the potential to be transformed by AI. For instance, to improve crop yields in the agricultural industry, self-driving tractors are using image recognition to identify weeds and spray herbicides on unwanted plants.

In healthcare, drug discovery is an area where generative AI could be extremely beneficial. Companies are seeing significant increases in the speed and efficiency of drug development, thanks to AI’s ability to analyse enormous datasets and recognise patterns. In an era of generative AI, pharmaceutical companies have the potential to make new medical discoveries and healthcare advancements that were previously unimaginable.

“Pharmaceutical companies have the potential to make new medical discoveries and healthcare advancements that were previously unimaginable.”

Increasingly, utility companies are embracing AI to help improve electricity infrastructure by enabling the development of ‘smart grids’, which will help optimise electricity distribution and consumption.

In the financial services industry, McKinsey suggests that banking is one of the industries that could see the greatest value from the technology^[3], particularly in the areas of risk assessment and fraud detection. The provision of services such as client reporting, customer service, and back-office operations could also become significantly more efficient with the use of generative AI. 

10 billion datapoints…and growing

At M&G, we are embracing technology and developing bespoke AI solutions to process data related to our own operations and investments.

The use of technology in the investment industry has grown exponentially over the past decade, in line with the rapid growth and evolution of data. There are a number of factors at play which are prompting a need for more data collection, and the subsequent adoption of tech-enabled solutions to process the data and make sense of it.

The changing regulatory backdrop means that we’ve seen an explosion in the volume of data related to our own operations and investments, whether it is Diversity and Inclusion (D&I) and climate data on our operations, or broader Environmental, Social and Governance (ESG) data on our investments. Asset managers have been realigning resources to comply with new regulations and disclosure requirements – both at a corporate and a product level – acquiring and analysing data and feeding this information into company-level and investment management systems and processes.

Meanwhile, in an increasingly competitive industry, the effective collection and use of data can make a real difference to the soundness of the investment process, as well as to generate comprehensive client reporting. Tech-enabled solutions can provide a richer array of inputs for investment analysis to enable differentiated insights and, ultimately, create better outcomes for our clients. It can also improve efficiencies and provide additional flexibility for our investment teams.

As active managers at M&G, the human element in the investment process will always be critical to our success. However, technology plays a pivotal role in helping us create innovative solutions and also scale up existing processes for greater efficiency. In essence, technology exists to serve a particular purpose – helping us translate data into actionable insights, more efficiently.

In our machine learning-based investment strategy, we use both AI and a human element. We have developed algorithms which form the core recommendation engine, and we overlay its conclusions with oversight from our fundamental equity team. The aim is to leverage the best of human and machine intelligence in the stock selection process. And in the same way as human knowledge grows, also the data size and compute power have increased over time – just more exponentially. As an example, when we first started employing this strategy in 2018, the underlying dataset comprised 1 billion datapoints covering 10 years of history. This is now over 10 billion datapoints covering more than 25 years of history, and growing every day.

Spectrum of AI opportunities

We believe natural language processing (NLP), whereby computers learn to understand and interpret human language, is just emerging as a powerful force in the global economy and the use cases will increase in line with improvements in processing power.

In our view, AI is a multi-year investment theme that has the potential to become a mainstream technology in the real economy. When we think about the spectrum of AI-related investment opportunities, we see three distinct categories: enablers, providers and beneficiaries.

AI enablers

Generative AI relies on a large amount of computational power, both to ‘train’ the LLMs and recognise the patterns in the data sets, and then to run the application and generate the responses.

Enablers are the companies that supply the key technology underlying AI advancements, such as computational power, data, high speed data communications, and sensors. These include semiconductor firms that produce the processors and chips used to perform high-speed data; systems companies that supply networking equipment; and telecoms firms that provide bandwidth to the network edge from the data centre and vice versa.

In our view, Nvidia is the most prominent enabler. The company provides most of the processors used to build neural networks to perform AI – this helps explain why Nvidia’s share price has soared more than 200% this year, pushing its market capitalisation over US$1tn^[4].

Without Nvidia, AI would arguably not be at the scale it is now – in the past 10 years, Nvidia has increased AI performance by 1 million-fold. In our view, Nvidia remains central to the future of AI as its CUDA^[5] programming language is the de facto standard for using Nvidia’s processors for AI. Demand for Nvidia’s chips is robust as tech firms across the world are seeking to build their own LLMs.

The company has a long-term vision and is constantly looking for new AI applications for its technology. In our view, the technology will only get better, driving further demand for Nvidia’s processors. As long as the company can continue to provide more capabilities, we expect Nvidia to continue growing at a rapid rate.

Queuing for computational resources

Given the increased computation power required to run LLMs, we believe the explosion of AI applications will translate into rising data centre demand. Companies are queuing to get access to more computational resources and increasing demand could lead to both higher revenue growth and better pricing power for data centre operators, in our view.

The pace of demand has also spurred cutting-edge research into ways to increase the speed and capacity of data transmission and storage. Faster data speeds are critical for AI performance as data needs to move quickly from the edge of the network to centralised data centres for processing and then the answer is sent back again. Amazon’s Alexa and Apple’s Siri are examples of how NLP is moved from the edge to the data centre and back out to the edge.

Optical connections can deliver data faster than electronic networks and are likely to be required to keep up with demand for ever greater speed of transmission. A number of companies have been investing in ‘photonics’ technology based on the use of light waves. Photonics is the science and technology of generating, controlling, and detecting photons for various applications such as telecommunications, medical imaging, and manufacturing, using lasers and optical components.

With data creation and transmission increasing exponentially, particularly with the wider adoption of AI, companies at the forefront of this new technology could be well-positioned to capture the demand growth.

AI providers

Another potential source of investment opportunities are the providers of AI services and products to end users. These companies typically, but not exclusively, belong to the software industry, which are enhancing their products with AI.

A notable provider, in our view, is Microsoft. The technology firm is a significant investor in OpenAI and is rapidly supplementing its software products with AI features such as Microsoft 365 Copilot. With a dominant market share in office productivity software, Microsoft has the opportunity to increase the power of its software and charge more for AI-integrated products. Other examples include human resources solutions firm, Workday, and creative software provider, Adobe, discussed above.

AI beneficiaries

Our third category is AI beneficiaries – firms that are increasing efficiencies by using AI in their operations. These are typically found outside the IT sector and span the automotive, healthcare and industrial industries.

US food and beverage firm PepsiCo is good example, in our view. The company is using AI in multiple areas to help grow its business, as well as improve operational efficiency. In manufacturing, PepsiCo is employing AI to maintain food standards. The company is using AI in product development, to gauge the potential success of new products; it also uses AI to determine shelf locations for its products, aimed at improving sales.

Over time, we anticipate that AI will be adopted by more and more companies, in every sector, as they seek to become more efficient and improve their productivity. As AI becomes more pervasive within the real economy, we would expect to see greater investment opportunities among the AI beneficiaries category in future.

“As investors, we focus on identifying the intersection between technological innovation and its implementation in business processes.”

As investors, we focus on identifying the intersection between technological innovation and its implementation in business processes. In our view, growing free cash flow (FCF), the money that is left after subtracting capital expenditures, is the most important factor in driving investment returns.

We believe that companies providing AI and/or using AI to improve growth and profit margins are likely to have strong FCF growth over the long term, which could provide a powerful boost to their investment performance.

Thriving and surviving

We have highlighted just a few examples of the impact generative AI could have on different sectors. There are plenty of other use cases, such as in education, manufacturing, and autonomous vehicles, and there will undoubtedly be many more as the technology continues to develop at pace.

Over the coming years, we believe that generative AI and its applications have the potential to be long-term growth drivers, boosting productivity in a range of tasks and transforming industries. In our view, this is an exciting time for innovation and competition – these are early days for the new technology, and it will be interesting to see how it develops, how firms harness it successfully, as well as observing the regulatory and security developments.

“We believe that generative AI and its applications have the potential to be long-term growth drivers, boosting productivity in a range of tasks and transforming industries.”

As ever, the start of a new disruptive market force will create winners and losers. We see plenty of attractive AI-related investment opportunities. But some companies may find themselves not being able to keep up with more AI-savvy competitors or may end up seeing the markets they operate in fall victim to AI’s many applications. On a multi-year basis, we believe that the implementation of AI, and generative AI in particular, will be but one determinant of those winners and losers.

 

[1] Reuters, “Chat GPT sets record for fastest-growing user base – analyst note”, Reuters.com, February 2023.

[2] McKinsey, “The economic potential of generative AI: The next productivity frontier”, mckinsey.com, June 2023.

[3] McKinsey, “The economic potential of generative AI: The next productivity frontier”, mckinsey.com, June 2023.

[4] Financial Times, “Enthusiastic Nvidia investors may need a reality check”, ft.com, July 2023.

[5] CUDA® is a parallel computing platform and programming model developed by NVIDIA for general computing on graphical processing units (GPUs). With CUDA, developers are able to dramatically speed up computing applications by harnessing the power of GPUs.