ARTIFICIAL intelligence will not replace individuals but people may be replaced by people who have a better understanding of how to use AI.
This was one of the key messages delivered to an audience of finance industry professionals and private clients at an event hosted by Quilter Cheviot last week.
At the briefing, titled Generative AI And What You Need To Do, Quilter Cheviot global technology analyst Ben Barringer answered questions from Jersey Finance deputy chief executive Amy Bryant and members of the audience on the development of AI, how businesses could apply it and the ethical considerations around its use.
Acknowledging that artificial intelligence had existed since the 1950s, Mr Barringer explained that it was the advance of cloud computing and the development of “huge data sets” that had facilitated generative AI.
“Generative AI relies on huge amounts of data,” he said. “In the case of ChatGPT, this information is fed into the system, which is then trained to answer questions and produce content based on that data. If, for example, you typed ‘the cat sat on the…’ and asked the tool to fill in the blank, it would provide the most probable answer based on the training and data it has received. Therefore, it is important to note that while it will give you the most likely answer, this may not be the correct one. This is why I remind people that, although it is good at some things, it is not a panacea.”
There were, however, four areas where Mr Barringer said generative AI could help businesses to boost their productivity.
“It is really good at content generation, summarisation, coding generation and semantic searches,” he said.
When asked by Mrs Bryant about the governance or oversight mechanisms which businesses needed to consider if using generative AI, Mr Barringer stressed the importance of a “clean and consistent data model”.
“This is a big challenge for most organisations,” he said. “Your AI is only as good as the data you give it and the patterns that it sees within that data. You also need to remember that the AI output may not be correct. Hallucinations do happen, so you need to check that output very carefully.”
Having focused on the potential uses and risks which AI presented to businesses, Mr Barringer then turned to investment opportunities.
“A precondition of generative AI is having data stored in the cloud, and the big players here are Amazon Web Services, Google Compute Engine and Microsoft Azure,” he said. “But there are also tremendous opportunities from companies which produce the components that go into those data centres. Companies such as Nvidia have seen an incredible share price performance, while firms such as TSMC, Samsung and Micron are also world-class technology companies, developing core components such as memory solutions and chips.
“We also look at companies which can infuse AI into their software and those, such as SAP and Salesforce, which are in a ‘sticky’ market, where it is difficult for people to move to competitors. Critically, these companies must also be large, with large amounts of data and knowledgeable AI scientists, giving them that all-important expertise.”
As questions turned to cyber security, the ethics around AI and the sustainability implications, Mr Barringer acknowledged that regulation was at an “early stage”.
“As with all forms of technology regulation, it is difficult because regulators are trying to develop laws for today but, by the time the legislation has been passed, the technology has moved on significantly,” he said.
“From a sustainability point of view, energy consumption is the biggest factor. In Dublin, which has one of the biggest data-centre footprints in Europe, 40% of the city’s energy is used to power these centres. While we have asked Microsoft whether it can deliver AI and still achieve its sustainability commitments, we will need to monitor this situation, and make sure that these big players stay true to their targets.”
Mr Barringer also highlighted the fact that AI was becoming a “new weapon of war”, with nation states creating AI research centres.
“This is not dissimilar from cyber warfare, something which we have seen in Russia’s invasion of Ukraine,” he said. “It is therefore vital that cyber security protects not just companies but also countries’ utilities and infrastructure.”
Bringing the topic back to the businesses in the room, Mr Barringer stressed that companies needed to “experiment with AI” to understand what it could do.
“You have to understand the tool and its capabilities, so that it becomes something we embed in the way we run our lives,” he said. “Human capital substitution is nothing new. Things like the printing press, the plough and the steam engine all revolutionised the way things were done, but they also showed how good technology is at performing more mundane tasks.
“Generative AI can get us to the start line quicker, so that we can automate more processes. But, as I say, it’s important to understand what it can and can’t do, so be curious and find ways of using it which allow you to carry out more value-added tasks.”
