Artificial Intelligence - Global Use cases

The release of Chat GPT 3, last year, showed the world’s population the benefits that artificial intelligence (AI) can bring to their daily lives. I showed my children how they can use it to help them in their homework, rightly or wrongly! The new version, Chat GPT 4, uses data and information from the internet, compiling 1 trillion parameters and 32,000 fine tuning iterations to answer questions and even write essays. It is basically a large language model that utilises artificial intelligence. Large language models solve language problems, such as text classification and generation, question answering and document summarisation. Other AI chatbots include Microsoft’s Bing Chat GPT, Google’s Bard and Baidu’s Ernie.

We have been investors in technology companies for decades and have invested in many companies that continue to develop their artificial intelligence capabilities. While the market is bidding up shares now, many companies have actually been working on this technology for over a decade.

I remember researching IBM’s Watson, which is one of the pioneers in artificial intelligence. Watson has been applied in industries such as healthcare to advance medical systems. Our largest position, AFYA, is applying artificial intelligence to their software to assist physicians in diagnosing diseases and recommending pharmaceuticals.

We have been investors in Baidu for a long time and have been watching them integrate artificial intelligence into their autonomous driving platform, Apollo, as well as their cloud software, Paddle Paddle. They have also been testing their Ernie chat bot in China this year. Baidu recently tested Ernie 3.5 against Chat GPT 4. It was ahead of GPT4 on admission and qualification tests to qualify as a lawyer but was behind questions developed by a group of US universities on science and humanities. It was ahead on 13,000 multiple choice questions covering fitty different subjects in Chinese.

Baidu has been improving their Paddle Paddle cloud software for years and have incorporated deep learning models into their own products. Now more than 4.8 million developers and 180,000 businesses worldwide use it for vision and natural language processing. Paddle Paddle has been sold to many transport entities in China, installing AI models to monitor traffic lights and improve traffic. It is used by recycling businesses to create object detection models for robotic garbage sorting. It is being used in agriculture to monitor crops, deploy drones to fight fires.

Baidu is also at the forefront of autonomous driving cars. Their Apollo Go, robo-taxi service, provided 660,000 rides in the first quarter of 2023. Their platform uses sensors to gather data about the surroundings, analyse the data using AI algorithms and then make real time decisions to steer the vehicle.

Databricks, an unlisted company, packaged up machine learning projects into an open-source software product. They are enabling companies with data in multiple different warehouses to access all their data and apply artificial intelligence to make decisions. Walmart is using databricks to apply artificial intelligence to their supply chain management.

We recently invested in Fluentdocs, which is a rules-based software solution, enabling industries like the legal profession to automate documents. They are integrating artificial intelligence into their product to enhance the automation process.

Artificial intelligence is being used to improve the advertising and design industries. We recently invested in Adobe, which is the leading provider of creative software with its Photoshop software. It is also an important player in marketing and document management tools. Adobe has enhanced its products with artificial intelligence and has been giving away its new tools on a new website called Firefly, enabling users to test their photo-editing tools, along with a service that can create images from a simple text command. For photo editing they are using their own platform, but for text they are leveraging ChatGPT. Companies will be able to either use Adobe’s content library or tweak Adobe’s model with their own content.

S4 Capital, a digital marketing and technology solutions provider, says that we are not far away from high-end film production generated from artificial intelligence. AI can be used to scale out photoshoots to get work done faster. It will reduce manual labour and enable more customisation of ads. It will also be used to personalise copywrite and emails from CRM systems. It will enable companies to provide their customers with live product design choices based on voice interaction with a machine, enabling you to design your own sneakers for example. Advertising groups that are at the forefront of this change will benefit, but the whole industry will see a reduction in its labour force and some businesses will lose business from these new efficiencies.

Some companies are just as focused as using artificial intelligence to cut out costs and improve their productivity as much as they are at generating new tools that will grow their revenue. Paypal has been leveraging AI in their risk and compliance divisions for years. They are currently spending a lot of time on testing it in their customer service and operations to provide smarter chats in self-service to fix issues, which could create additional cost savings. They are also looking to apply AI models to their large amounts of data to better profile their customer base.

While some companies and industries will benefit from the use of artificial intelligence to improve productivity and create new services and offerings, other companies and industries will see some of their revenue streams and offerings made redundant. The world is constantly evolving, but you can take heart in knowing that this has been a development that we have been witnessing for the past decade and it will continue to play out over the next decade.