Without a doubt, generative artificial intelligence (AI) is transforming businesses worldwide. But even as it moves out of testing and into production, the rate of progress is not uniform.
As a rule, European countries such as Ireland lag slightly behind the US when it comes to adoption of the technology. So while the US has made significant strides, European organisations seeking to catch up must overcome regulatory hurdles and invest in data infrastructure.
According to a 2023 McKinsey survey on the state of AI, 40 per cent of American companies reported adopting gen AI in at least one business function, whereas around 30 per cent of surveyed European companies said the same.
This is not entirely surprising, given that the major developments in generative AI were spawned at US companies such as OpenAI and Google. However, there is a second explanation for the US lead, said Lee Sexton, client technical director at the global IT-enabled infrastructure, cloud and security solutions provider Presidio: regulation.
“Things like GDPR [the EU’s General Data Protection Regulation] make a lot of things trickier, and there are a number of clients, people in the medical sector, who are doing more things in the US than their equivalents in Europe,” he said.
“Privacy and compliance are big topics at the moment and, from a European point of view, you have not only GDPR but also some other regulations such as the European Data Act. The reality is that there is a degree of policing coming in, and that does make sense when you look at what can be done with these tools.”
In a lot of use cases, it’s about building a pyramid up to AI, starting with getting the data into order.
Sexton said this did not mean that the pace of adoption was not accelerating in Europe. However, serious integration of AI into operations requires businesses to have a data strategy ensuring not only compliance with the law but also that useful data is available.
“I’d say it depends on what you’re doing. You’ve got certain industries where gen AI is definitely going to help. Also, people who learn to use generative AI and develop skills in prompting will get a proactive increase, but for really deep business-critical applications there is a need for data management and data cleaning.
“To some degree, that’s a big conversation. In a lot of use cases, it’s about building a pyramid up to AI, starting with getting the data into order. However, in other industries they can jump directly into generative AI.”
One example where rapid deployment was possible, he said, was in customer service. Contact centre duties can be significantly lightened by the addition of generative AI as there are significant knowledge bases for the AI models to work from.
“For customer-facing things, you can do it more quickly. If you take a call centre, or FAQ bots that exist today, they have the ability to use the learned information from an agent-base of knowledge,” he said.
This was also a good example, he said, of how AI’s role was to assist and augment, not to replace: generative AI could be trained to answer straightforward questions, meaning staff could deal with the more complicated queries.
“It’s in cases like this where the productivity boost comes in. It’s not going to replace people’s jobs; it’s augmenting them. People are queuing in a contact centre because either it’s not cost effective to hire people or they aren’t able to hire anyone.”
These are not theoretical applications either. Presidio has clients who are using generative AI, and other forms of AI, today.
“It’s out in the wild already, typically in very discreet user cases, and more in the contact centre space,” Sexton said.
“We also have a lot of clients using machine learning [ML] and data to make decisions. Effectively, what they’re trying to do is make a decision on what’s next. For example, in trading you want to ask things like: ‘is it the right price?’ or ‘has the futures market bottomed-out?’,” he said.
Indeed, trading and investing is an area where technology such as ML is likely to have a major impact, allowing retail traders to emulate the quantitative trading strategies of hedge funds such as Renaissance Technologies, where a team of PhD mathematicians pioneered algorithmic trading.
“In those spaces, the ability to make those decisions will become cheaper. Ten years from now it will all be simpler, and perhaps you won’t need a team of mathematicians,” he said.