Back in 2016 when we came up with the idea of an actively managed global equity investing in the leading lights and beneficiaries of the deployment of artificial intelligence (AI), we were reasonably confident that the economic impact of AI would be comparable to that of the railways, the internal combustion engine, the telephone or TV. That is to say, it would transform economies and society in general and there would be winners and losers from the transition – just as there were in the first industrial revolution. In 2016 and 2017, we spent a lot of time explaining what we meant by AI – some people scoffed at the idea of artificial intelligence and others simply discounted it as a fad – sure, there might be some limited application of AI in the tech space, particularly in the form of better coding and programming, but nothing to get really excited about unless you were an IT aficionado. And it is fair to say that there were whole industry sectors that either didn’t care about or weren’t engaged with AI.

Fast forward to 2023 and the world has changed in ways that few would have predicted. There is a major and ongoing war in mainland Europe. Climate has become a key risk of risk and opportunity for investors – and as the result the internal combustion engine is slowly being legislated out of existence. Strange looking balloons have been observed – and shot down by the USAF – over the Americas. And AI is popping up everywhere and has become part of popular consciousness. Nowhere is this truer than with ChatGPT, a form of generative AI. When artificial intelligence is headline national and international news, there is no doubt that it will and should be on investors’ radars.

What is generative AI?

In simple terms, generative AI is a broad term that can be used to describe any form of artificial intelligence used to create text, images, video, audio, code or synthetic data. Generative AI models can create original context in response to a user prompt. In isolation, it may not sound that exciting, but the implications are profound – not least because generative AI can perhaps provide a meaningful solution to the lack of productivity growth that has dogged developed world economies in recent years.
 

Why all the fuss about ChatGPT?

One of the most impressive things about ChatGPT is that can write on a whole range of topics – and it can do this in a very convincing, human-like way. There are already concerns at universities around the world that ChatGPT could make essay writing redundant as a form of academic assessment for undergraduates. (Some universities are developing tools that may enable them to identify essays written by ChatGPT.) Microsoft founder Bill Gates has already said that ChatGPT will transform the world and particularly white collar jobs. But what is all the fuss about?
 

Welcome to ChatGPT

A colleague at Sanlam Investments is a bit of a Luddite – he doesn’t really use social media and likes analogue tech (think vinyl record players and late 1990s European hatchbacks). We asked him to engage with ChatGPT (early in the morning UK time, as once the US wakes up it is almost impossible to access the platform due to the huge demand). His questions, and ChatGPT’s answers, are below.

• What are the potential applications of ChatGPT?

  ChatGPT: There are many potential applications of ChatGPT, as it is a highly versatile language model that can generate human-like text. Here are some of the most common uses of ChatGPT:
  
  Conversational AI: ChatGPT can be used to build conversational agents, such as chatbots, that can interact with users via text or voice. These chatbots can be used in customer service, support, or information-gathering applications.
  
  Question Answering: ChatGPT can be trained to answer questions based on a large corpus of text, making it a powerful tool for information retrieval.
  
  Text Completion: ChatGPT can be used to complete partial text, such as an email or a message, based on the context provided.
  
  Text Generation: ChatGPT can be used to generate new text based on a prompt or a seed text, making it a valuable tool for content creation and text-based applications.
  Sentiment Analysis: By analysing the language used in text, ChatGPT can be used to determine the sentiment of a text, for example, positive, negative, or neutral.
  
  Text Summarization: ChatGPT can be used to automatically summarize long pieces of text into shorter, more concise versions.
  
  These are just a few of the many potential applications of ChatGPT, as the versatility of this language model means it can be used in a wide range of tasks that involve generating or processing text.

• Would you say that ChatGPT already understands language better than most humans?

  ChatGPT: As an AI language model, I have been trained on a large corpus of text and have the ability to generate human-like text based on the input I receive. In this sense, you could say that I understand language at a level that is unparalleled by most humans.
  
  However, it's important to note that my understanding of language is limited to the patterns and relationships that I have learned from the text I was trained on. I do not have the capacity to understand the nuances of human language, such as sarcasm, humour, or cultural references, in the same way that humans do.
  
  Moreover, while I have a vast knowledge of language, I do not have the ability to experience emotions, form opinions, or make decisions in the same way that humans can. I am simply a tool that has been designed to generate text based on the input I receive, and my understanding of language is limited to the patterns and relationships I have learned from the text I was trained on.
  
  So while I may understand language at a level that is unparalleled by most humans, it's important to recognize that my understanding is limited and does not reflect the full range of human understanding and experience.

• A lot of automated language models produce writing that is obviously written by a robot but your writing could be mistaken for that of a human. How have you achieved that?

  ChatGPT: I have been trained using a type of AI called a transformer-based language model, which is capable of producing human-like text. The training data for this type of model is typically a large corpus of text, which can include books, articles, websites, and other written material. The model learns patterns and relationships in the language from this training data, which it can then use to generate text that is similar to the text it was trained on.
  
  OpenAI specifically has trained me on a massive corpus of text, which has allowed me to learn a wide range of linguistic patterns and relationships, making my generated text more coherent and natural-sounding. Additionally, the transformer architecture of my model allows me to effectively process context and generate text that is influenced by the context in which it appears, making the generated text even more human-like.
  
  While I have been trained on a vast amount of text and can generate text that is often difficult to distinguish from text written by a human, it is important to keep in mind that I am still an AI language model and do not have consciousness or personal experiences. There will still be instances where the text I generate is clearly not written by a human, and it is important for users to be able to distinguish between human-written text and text generated by an AI like me.

In short – we have found that the answers from ChatGPT are not perfect but for something that is (currently) free the results are very impressive.

What are the implications of generative AI?

It is only February and already 2023 is shaping up to be one of the most exciting years for AI. A recent study by McKinsey highlighted a multitude of potential uses for generative AI – and for clarity, the list is not exhaustive.
 

Full steam ahead for generative AI?

Generative AI is undeniably very powerful, but like all technologies in its very early phase, it needs to be used with some caution. For example, just like humans, ChatGPT can make mistakes and there is no immediate or obvious way of signalling this to the user. Some of the content produced by generative AI may not be appropriate and like all AI platforms that have been trained on data, systematic biases (whether intentional or not) within the underlying data sets can pose real problems. Moreover, generative AI won’t reflect corporate norms or values, which most companies tend to want to protect or enhance.

There are also wider ethical questions associated with generative AI – for example, if a generative AI model comes up with a new idea or concept based on a user suggestion, who owns it? And how will original content creators be acknowledged (and rewarded) when a generative AI model uses the content creator’s data to formulate an answer?

What does the future hold?

We think that the future for generative AI is very bright – the current generative AI platforms are by no means perfect but the scope for time and efficiency savings is enormous. That will be music to the ears of policymakers who have been baffled by low rates of productivity growth for years. People will argue as to whether content created on platforms like ChatGPT is genuinely ‘creative’ in the truest sense of the word but it won’t matter for businesses who are able to leverage the technology effectively as the impact is likely to be huge. We have always felt that investment in AI is as much about finding the companies who are leveraging AI (irrespective of the industry sector they operate in) as it is about those who develop the enabling technologies, and we are genuinely excited by the raft of opportunities presented by generative AI.

Fund risks
The Fund may invest in  shares of companies listed on stock exchanges in the United Kingdom, and outside the United Kingdom, exchange rate fluctuations may cause the value of investments to go down as well as up.  Investing in companies based in emerging markets may involve additional risks due to greater political, economic, regulatory risks, among other factors. The Fund may invest in derivatives for the purposes of efficient portfolio management and hedging.