Discussion with

Alison porter and SSarah de Lagarde

Sarah de Lagarde (SDL): Is AI just another hyped up technology theme, feeding into higher prices and company valuations? Or is there more to it?

Alison Porter (AP): It’s definitely more than a technology theme. We consider it as a compute wave like the mainframe computer, the personal computer (PC), and the internet. A few factors distinguish compute waves from themes. First, while the catalyst is always a particular product combination, compute waves are much broader. They touch every single layer of technology, including the computers we use, servers, and storage, all the infrastructure. Second, they often come to prominence after a crisis, out of the need for productivity gains, such as during COVID-19. And third, they are long-term in nature as lots of investment is needed.

SDL: What do you think will be the main applications of AI?

AP: Transportation is a key area. It is still early days, but one very useful application is active safety, to reduce the number of accidents as around more than 90% of accidents are caused by human error.¹

Another area which is a real focus is healthcare. On average, a hospital produces about 50 petabytes of data per year.² That’s a massive amount of data – since one petabyte is equivalent to 20 million tall filing cabinets of documents.³

In the search for a COVID vaccine, the access to data lakes (repositories of stored data) and rapid machine learning helped drugmaker Moderna find quicker vaccine solutions.

We also think AI can be particularly useful for workflow management, improvement of patient throughput, and the early detection of diseases. We’re seeing that being successfully applied in areas like radiography and robotic surgery.

SDL: That absolutely resonates with me because I had a terrible accident on the London Underground. During the accident investigation I came across technology applications that are essentially AI-powered algorithms that can read CCTV footage in real time. For humans, it’s impossible to view 2,000 camera feeds at once, but AI can do that, analyse the footage and spot irregularities. Sensors can spot if someone falls onto the tracks. So, lots of ways of improving safety and reducing the number of accidents.

But I also like the fact that you spotlighted the healthcare sector; the pace of innovation there is incredible. I’ve been fitted with a bionic arm that is powered by artificial intelligence. And I always think, things can only get better because the technology is only going to improve – I find that quite reassuring. So, I guess we both can see huge benefits to leveraging AI.

However, there is one big question that makes the headlines: Are the risks in AI manageable? Is there such a thing as responsible AI?

AP: We think the risks can be manageable, and there is a path to responsible AI. Your story is incredibly inspiring, and a great example of the many benefits that AI can bring. It highlights why innovation is so essential, and that AI is already being put to good use.

There are, however, several risks. Discrimination and bias, for example, in the use of facial recognition technology mistakenly identifying criminals. Or the early issues around Google’s Gemini AI image generation application, which overcompensated for bias by generating historically-inaccurate images of people. There needs to be a balance between avoiding bias and stereotypes and ensuring accuracy.

Another risk is transparency. We want to know when AI is being used. Are we seeing a real or an AI image? And knowing when human interaction is needed for decision-making. To be clear, we’re referring to Generative AI in terms of new content generation, not a full-blown shift to artificial general intelligence.

We believe active investors have an important role to play in shaping how companies consider developing and reporting on Generative AI use over time.

SDL: Can you explain the difference between AI applications today and artificial general intelligence.

AP: Sure. AI today is able to process vast amounts of data and generate new content, often being very creative. So we’re very much talking about AI being an assistant, a “copilot” to human decision-making as Generative AI doesn’t understand the content it creates at a deeper, more human level. Artificial general intelligence (AGI) is a further leap ahead as it more closely mimics cognitive abilities of humans, not only replicating or assisting actions but grasping the intricacies and contexts of actions generated.

A key role for us as active managers is to question and understand companies’ thinking and approach to responsible AI, as well as their reporting on it. Governments will play a big part here and there are already several different models and approaches emerging in different regions, from disclosure rules to oversight and licensing controls. As investors, we’re well positioned to shape companies’ thinking on their approach to transparency, to ask questions on their oversight frameworks about who is responsible and accountable within those companies.

So, I think working together, alongside regulation, we can build a real trust network. We recently carried out a proprietary survey with the University of Liverpool, looking at demographics.

Encouragingly, we found that young people see the opportunities of AI and not just the risks. That bodes really well for managing the risks and developing responsible AI for a better future.

SDL: Listening to some of our colleagues in the media, there are some concerns that there could be a doom- and-gloom scenario if we don’t monitor developments closely. But for me, it feels like there is a seismic shift in technology. We’re standing at the precipice of a new dawn.

Source: Janus Henderson Investors

Footnotes

¹ Source: European Commission, Road safety: Commission welcomes agreement on new EU rules to save lives, 26 March.

² Source: World Economic Forum, “4 ways data is improving healthcare”, December.

³ Source: com “What is a petabyte?”, June 2024.

References made to individual securities do not constitute a recommendation to buy, sell or hold any security, and should not be assumed to be profitable.