AI paradise awaits the telco-friendly regulator, claim Liberty Global and EYAI paradise awaits the telco-friendly regulator, claim Liberty Global and EY

It makes eight recommendations, that include assessing AI's sustainability impact to identify avenues for improvement, and prioritising the development of AI-driven network optimisation.

It also recommends the swift transition to renewable energy and the adoption of circular economy practices. Organisations should also foster a culture and conduct training to ensure that they have the values and skills needed to drive innovation in tandem with sustainability.

Liberty and EY also recommend the development of AI governance frameworks that integrate sustainability considerations, as well as industry-wide standards for sustainable AI, including common metrics for measuring environmental impact.

Do all that, and AI might just help rather than hinder the industry's net zero efforts, the report claims.

"As data traffic continues to grow exponentially, we are committed to playing our part in reducing emissions throughout our networks and operations," said Liberty CEO Mike Fries. "We're excited about the potential of AI to help us do this – and we're already seeing the benefits in areas such as the optimisation of energy use in our networks and making our network builds more efficient."

The report's credibility is undermined somewhat by the fact that chunks of it read like a barely-concealed attempt to lobby for a telco-friendly regulatory environment.

Maybe this is too cynical a take, but when Ericsson and Nokia organise a joint summit to present their case to the new EU Commission, and when Ecta is sending delegations to ensure its voice is also heard, it is hard not to believe that Liberty is really keen for this report to land on a few important desks in Brussels.

The report posits four possible futures, ranging from the wildly optimistic to the apocalyptically terrible. Each one is determined by policies that either support or undermine AI to varying degrees, and the knock-on effects this could have on adoption by the industry, society at large, and by extension the environment.

The ideal scenario, dubbed 'Transform', describes a world where the industry favours a targeted rather than blanket approach to AI adoption, leading to a utopia where the average telco reports an 80% reduction in network energy consumption, despite carrying 100 times more data traffic.

Unsurprisingly, this future is ushered in largely thanks to the implementation of pro-investment regulations as laid out by the Draghi Report, the recommendations of which include allowing cross-border mega-mergers and harmonised spectrum rules, among others.

At the opposite end of the spectrum is 'Collapse', where the unabated employment of resource-intensive AI workloads in every facet of life has led to a 150% increase in energy use by telco networks.

In this scenario, unscrupulous tech giants – the ones that some telcos claim have been enjoying a free ride on their networks for far too long – have thwarted attempts to regulate AI energy consumption.

'Constraint' sits somewhere in the middle. This is where heavy-handed regulatory intervention has led to transparency, fairness and data protection, but has also limited the scope of AI performance, making it harder to realise the potential energy-saving benefits the technology has to offer. As such, telco network energy consumption is only 30% lower, and the volume of data they carry has only increased three-fold.

Striking a more optimistic tone is 'Growth'. This is the future that Liberty and EY think the industry is headed for.

The picture's not entirely rosy: in this scenario, telco network energy use actually increases, albeit by a modest 10%. However, scope 1 and 2 emissions have been reduced by 100%, partly due to networks being able to predict and prevent 99.9% of faults, drastically reducing the need for energy-intensive manual intervention and field visits.

AI also enables network traffic to be routed more efficiently to handle the predicted 50x increase in load, and data centres have used AI to reduce energy consumption by 80%.

This future is supported by forward-thinking regulatory frameworks that encourage AI adoption while mitigating potential risks to critical national infrastructure.

"While the scenarios we present are speculative, they highlight the strategic potential of AI in the telco sector, said Harvey Lewis, partner for AI at EY, and the author of the report. "Success will depend on network collaboration across Europe, investment in skills and infrastructure to create more competition, and a holistic approach to AI adoption that balances immediate efficiency gains with long-term sustainability impacts."

There have been a few notable developments in the real world lately when it comes to both measuring AI energy consumption and using AI to improve energy efficiency.

For example, BT in October updated its Carbon Network Dashboard – which enables businesses to track the energy consumption of their network and data centres in real time – to include AI applications.

And back in June, Three UK and Ericsson began using AI and machine learning to automatically power down radio components during periods of inactivity.

What Liberty Global, EY, and the whole telecoms industry – if not the world – hopes is that these practices become a normal part of day-to-day operations, rather than isolated occurrences.

And for Liberty and EY in particular, the right regulatory framework could make all the difference.