The foundational principle of net neutrality, which mandates that internet service providers treat all data traffic equally, is facing renewed scrutiny in the UK as the widespread adoption of artificial intelligence (AI) dramatically increases internet usage. This long-standing concept, credited with fostering immense growth across online services, is now being challenged by major telecom operators who argue that the current system is unsustainable.
For decades, internet carriers – including broadband and mobile providers – have been prohibited from discriminating against any online content or service, regardless of its origin or the volume of data it consumes. This has allowed even multi-billion-pound companies like Google, Netflix, and Meta to access the internet on the same terms as individual users. However, the chief executive of BT’s consumer business, Marc Allera, highlighted in 2022 that a small number of firms are responsible for a disproportionately large share of network traffic. He drew parallels to other industries, suggesting that heavier usage typically incurs greater costs, a principle not applied under current net neutrality rules.
The UK’s media and telecoms regulator, Ofcom, conducted a review of its net neutrality guidelines in 2022. While it briefly explored the possibility of allowing telcos to charge high-usage businesses more, Ofcom ultimately decided against it. The regulator stated at the time that there was insufficient evidence to justify such a charging regime, noting that any significant shift would ultimately require a decision from the government and Parliament.
However, the technological landscape has evolved rapidly since Ofcom's review. The mainstream adoption and skyrocketing usage of AI have introduced unprecedented demands on internet infrastructure. Hyperscale companies are projected to invest around £600 billion ($750 billion) in capital expenditure this year alone to support the data and computational needs of advanced AI models. Furthermore, the rise of localised AI development, with individuals building systems using open-source models, is expected to further strain networks by generating substantial additional data traffic.
Industry experts suggest that AI-related traffic could soon be orders of magnitude greater than any other form of internet usage. This raises fundamental questions about the continued viability of treating all traffic equally, especially if networks become predominantly portals for AI applications. Telecom executives are increasingly vocal about the need for a re-evaluation, suggesting that the current rules may need to adapt to this new reality.
Any move to dismantle or significantly alter net neutrality rules would be a complex undertaking, given the numerous vested interests and legitimate arguments for its protection. However, the increasing strain on network infrastructure from AI's insatiable demand for data is intensifying pressure on policymakers to consider whether a new approach to internet traffic management, potentially involving differentiated charging, is now necessary.