Water companies have positive story to tell on AI, says expert

Chris Martin of Pinsent Masons said the ‘Stream’ and ‘River Deep Mountain AI’ projects demonstrate the commitment of water companies to applying data and digital technology to achieve improved operational and environmental outcomes.

Martin was commenting after water companies in England recently came in for criticism over their use of AI and other digital technologies relative to counterparts in other parts of the world.

“Undoubtedly data, AI and indeed other digital technologies, tools and digital twins are going to play an important role in the modernisation and operation of water infrastructure and the industry as a whole in the UK,” said Martin. “There are significant opportunities to harness computing power, AI, and unlock value from data to help optimise how the sector operates. Indeed, the sector is already taking steps to embrace that – with the Stream and River Deep Mountain AI projects prime examples.”

Through the Stream project, water utilities from across the UK make a wide range of data available to access via an open data platform. Stream facilitates access to both open datasets and secures sharing of trusted data too and supports the publication of AI-ready data in line with best practice frameworks. The data includes information on drinking water quality, water consumption, discharges of storm overflows, reservoir levels, sewer flooding, and leakages and fixes. The project has received funding from Ofwat, water regulator in England. Pinsent Masons advises on legal and governance aspects of the project.

“Part of the driver for the project is an eagerness from water companies themselves to understand how they can use data to tackle the systemic issues that the industry faces, particularly around high-profile issues like leakages, pollution, and storm overflows,” said Martin. “The benefit of making the data available on an open data platform is that it enables third parties, such as technology providers and data-driven innovation businesses to access the data and potentially identify solutions to the challenges that industry is facing. This is where the power of AI has great potential – AI is only useful if it has high-quality datasets to feed on, learn from, consume and ingest.”

The River Deep Mountain AI project tested how AI and other digital technologies such as machine learning and remote sensors could be used to better anticipate and manage environmental risks in water courses compared to traditional practices that have “depended heavily on physical sampling, delayed laboratory testing, and fragmented datasets”, according to a project summary. The project was successful in demonstrating the potential of AI and remote sensing to “improve how the sector monitor and manage waterbody health”, has led to refined datasets being made available via Stream, and, among other things, has already led to “improved ability for water quality estimations, reduced reliance on manual sampling, enhanced multi-organisation collaboration, and stronger evidence for policy and investment decisions”, the summary said.

Martin said: “These projects have been driven by water companies themselves but align with the enhanced regulatory expectations on them around things like improving openness and transparency, embracing innovation and collaboration, and generally improving the performance – and public perception – of the sector. Encouragingly, the lessons learned around matters such as data quality and governance provide a strong foundation for the industry to push on to maximise the potential of AI and data, and digital technologies more generally, to support in managing water infrastructure in a more effective way.”

Last year, Pinsent Masons undertook a survey of its global clients on the impact of AI on the built environment, in partnership with Bentley Systems, Mott MacDonald, and Turner & Townsend. According to infrastructure expert Graham Robinson of Pinsent Masons, the findings show that the overwhelming majority of clients see a significant shift in their business models on the horizon from the use of AI.

Robinson said: “Productivity in the construction industry has been stagnant for the last two decades – AI is expected to unlock significant productivity gains, especially if linked to an industrialised construction approach.”

“Significant barriers to the effective use of AI for infrastructure construction exist, including the lack of common project data, digital skills sets and leadership, governance, and risk management,” he said.

Pinsent Masons hosted a series of client roundtables to discuss the findings of the survey with more than 100 clients in London, Madrid, Dubai, Abu Dhabi, Riyadh, Hong Kong and Singapore. Robinson said: “The discussions make clear that AI readiness is a key issue, and navigating fast changing regulatory environments provides challenges, but that once a tipping point is reached, the use of AI in infrastructure and the built environment is expected to accelerate at pace.”

Nils Rauer, Frankfurt-based expert in AI regulation at Pinsent Masons, recently spoke at the Spring Conference of EIC, the European International Contractors federation organised by SEOPAN in Málaga. He focused on the identification of prospering AI use cases in the construction industry, the importance of AI literacy, and how human oversight can work in practice.