The overall theme at this year’s Mobile World Congress was ‘intelligent connectivity’. Described by the event’s organisers as the ‘beginning of a new era of highly contextualised and personalised experiences’, it’s a theme that’s as equally applicable to service providers as to their customers, and one that will be increasingly enabled by the adoption of artificial intelligence (AI) and machine learning (ML) technology.

Once the sole province of future-gazers, AI has slowly but steadily progressed to become a critical part of our everyday lives. However, while virtually every device launched at MWC last year featured an element of AI or ML, the general opinion was that the industry was only just beginning to explore the technology’s potential. But, as the telecoms world assembles in Barcelona once again, we now appear to be on the verge of a ‘profound technological revolution’.

Use cases and opportunities

The advent of 5G has opened the door to several use cases in which AI plays an integral part. Edge intelligence, intelligent network slicing, and predictive digital twins for example, will all be greatly enabled by 5G, and all rely on AI to some degree.

Edge computing is on the rise. In order to deliver the ultra-low latency promised by 5G, compute resources are being moved to the network edge, reducing the distance that data is required to travel from where it is generated. A serverless architecture at the edge allows for greater scalability and availability than would be afforded by edge-based servers, while the application of AI-powered analytics enables edge intelligence, the level of which was only previously available in on-premise or cloud-based datacentres.

Network slicing, another function enabled by 5G, allows an operator’s business customers to tailor their connectivity and data processing capabilities to their specific requirements. The addition of AI, however, automates the process, thus optimising its efficiency and increasing the opportunities for operators to address the evolving needs of their customers. Indeed, according to the GSMA, network slicing open up a revenue opportunity worth around $300 billion by 2025.

5G will provide the connective fabric for a wealth of connected IoT sensors, the data from which can be used to create digital twins – digital replicas of physical assets, processes and systems. Predictive analytics then help organisations make sense of this vast volume of data, turning information into insight. Predictive digital twins can shape the future of a business. According to Deloitte, “digital twins can profoundly enhance an enterprise’s ability to make proactive, data-driven decisions, increasing efficiency and avoiding potential issues. And they can make it possible to “experiment with the future” by exploring what-if scenarios safely and economically.”

The future is closer than we think

We’re – fortunately – a long way from approaching Terminator or HAL 9000 levels of artificial intelligence. We’re starting to see the emergence of AI-driven applications, however, that until recently were the stuff of science fiction.

Autonomous robots, drones and vehicles, for example, use AI and ML to automate functions previously performed by humans. The manufacturing industry is being transformed by ‘smart factories’, in which collaborative robots, or cobots, are trained to perform a multitude of tasks rather than carrying out one repetitive routine. And self-driving cars, expected to hit the roads this year, rely on AI and ML to continually process the data they need to learn how to drive and perform on the road as a human would.

Extended reality, or XR, a mixture of real and virtual environments and interactions between humans and machines, is set to radically transform the way in which we interact with media. The hope is that, in time, people will not distinguish between AR, VR or video, and will seamlessly move between the media that best suits them at any given time.

Finally, AI will intersect with quantum computing, creating the ability to present multiple solutions to highly complex problems simultaneously. The potential of this ‘quantum machine learning’ will be hugely disruptive, enabling AI to more efficiently perform complex tasks in a human-like way, and allowing robots to make optimised decision on a given situation in real time.

The future is boring

These are all exciting prospects, of course, and updates on their progress will undoubtedly draw crowds at events such as MWC over the coming years. As the technology continues to evolve and develop, however, we must proceed with caution. Despite fears of AI-enabled Terminators, John Giannandrea, Apple’s SVP of machine learning and AI strategy, suggests we forget killer robots – bias is the real AI danger’.

Although AI is viewed as intelligence demonstrated by machines, the algorithms on which it runs, and the data sets used to train those algorithms, are all created by humans. As a result, they can be tainted, containing implicit racial, gender or ideological biases, clearly unhealthy for any applications supported by AI. It’s important, therefore, when developing AI systems to maintain an awareness of bias, and use quality, transparent data at all times.

AI is clearly a key pillar of ‘intelligent connectivity’, with the potential to transform our lives. For the time being, though, the AI applications that will make the most difference to our lives are going to be relatively dull when compared to a future of autonomous drones, quantum computing and widespread XR. And there’s nothing wrong with that. When it comes to automating and optimising their networks, operators are already realising that it’s the boring, background solutions that make the biggest difference. So, while there was a lot of talk about the bright future of AI at MWC – with dancing robots and automated baristas- it’s the little things we should really be paying attention to.