Daniel Brown

As part of Leeds Digital Festival I recently attended an event hosted by Harvey Nash and the University of Leeds discussing the compute and energy requirements of AI, and the impact this is having on both our consumption of carbon-based energy and the local and global impact of the huge data centres (or ‘AI factories’) being built to facilitate AI’s continued growth. The event was a screening of David Savage's new documentary ‘Data’s Untold Story: The AI Energy Crisis’, followed by a panel discussion. I was on this panel alongside Vinny Leach and Arash Rabbani and our discussion was facilitated by Charlie Deacon. I was really pleased to get the opportunity to both see the documentary and be a part of the discussion afterwards with such an engaged audience. We are witnessing a massive growth in the development and use of AI right now, in particular generative AI. The large language models (LLMs) and other foundation models that underpin GenAI are some of the most energy hungry forms of AI – the energy taken to train OpenAI’s GPT-4 model for example is estimated to be 50 gigawatt hours, even greater for the recently released GPT-5. And it takes energy every time we prompt GenAI – Nature reports that search using GenAI takes up to 23-30 times the amount of energy as traditional search. It’s also energy intensive to simply store data: the documentary states that around 60% of data stored today is single use – there is no point in storing it but it’s preserved all the same. It’s important to recognise that a multitude of different machine learning methods can be used to develop AI. Many of these methods are no way near as energy hungry as generative AI and they all have different merits in terms of application and the benefits they can bring. We discussed this on the panel, along with how the growing awareness of AI’s energy costs will lead to the reporting of energy usage alongside the use of AI, and how minimising the storage of rarely (or never) used data would have an immediate effect. continued in comments...

44

6 Comments

My latest post, 'Applied Knowledge Graph vs Knowledge Graph Layer: Practical Insights from a Labelled Property Graph Perspective', focuses on the topic of the difference between Applied Knowledge Graph (AKG) and Knowledge Graph Layer (KGL) from the prism of an LPG practitioner: * An Applied Knowledge Graph (AKG) is your tactical tool. It is solution-oriented and designed to solve concrete business problems, delivering direct value for specific workflows such as GraphRAG, risk analysis and compliance monitoring. Think of it as the 'applied kitchen', built for immediate impact. * A Knowledge Graph Layer (KGL), on the other hand, is your strategic foundation. Its purpose is integration and governance, creating a unified semantic fabric that makes all your enterprise data speak the same language. This is your 'enterprise pantry', ensuring long-term consistency and interoperability. The post explores how the flexible LPG model can serve as a practical foundation for both of these things, enabling you to start small and scale up as required. Full article: https://lnkd.in/eNENFpNG #KnowledgeGraph #KG #AppliedKnowledgeGraph #AKG #KnowledgeGraphLayer #KGL #GraphRAG #AI #EnterpriseAI #DataArchitecture #LPG #LabelledPropertyGraph #DataGovernance #SemanticLayer #DataIntegration

66

24 Comments

Having talked to data people across Scotland over the last few weeks, there is one massive thing that is coming through in those conversations .... The work that The Data Lab - Innovation Centre is doing in the data community in Scotland has been giving people new opportunities and support to work in data. The Data Lab Academy are providing workshops, structured programmes and funding scholarships for Masters courses in data, analytics, computer science, AI etc, and that is filtering through to the job market. This has resulted in a growing population of data people in Scotland with real skills, real opportunities and real experience. Well done and long may it continue!!

37

Open Response to the AI Opportunities Action Plan The DataKirk, Scotland’s grassroots organisation for promoting data literacy and analytics training, and digital inclusion, welcomes the UK Government’s announcement of the £14 billion AI Opportunities Action Plan. The plan outlines a comprehensive strategy for accelerating AI adoption across the UK, fostering innovation, and cultivating a skilled workforce to drive economic growth and improve public services. At the DataKirk, we are committed to ensure no one is left behind in the UK AI ecosystem. Through programmes such as our upcoming #GEN-AI AfterSchool Club, we aim to inspire and equip people, particularly those from minority, low-income, and migratory backgrounds with the skills and confidence to actively participate in the AI economy. Initiatives like these bridge the gap between policy aspirations and grassroots action, ensuring equitable access to the opportunities AI offers. By working together, we can ensure that this technology is used to benefit society, create a more prosperous future for all, and solidify the UK's position as a global leader in AI innovation. Richard Lochhead Manira Ahmad Albert King Abel Aboh Linda Scott Susan Lechelt Kasia Kokowska (MCIPR) Louise Drumm Tom Kane Maxwell Chipulu Tom Wilkinson Steph Wright Richard Edwards Carolyne Thomson Heather Thomson DataKirk Chuks Ododo Odunayo Ogunsanwo Azeem Adeyemi Nimat Jimoh-Umaru The National Robotarium Stewart Miller Peter Kyle Keir Starmer Kameron Bowstead

38

9 Comments

There’s a lot of debate about why data teams don’t always deliver the value they should. I think this is part of the reason. We were discussing process mapping as part of our transformation programme, thinking through how and when data would be captured within redesigned journeys. It became clear we needed to spend more time agreeing what the data is, what it means, and how it will be defined and used. In fast-paced delivery, it’s easy to focus on user flows, screens and forms, and assume the data will take care of itself. But without a clear, shared understanding of the data model, we risk creating processes that look good but fall short when it comes to reporting, automation or insight. Somewhere along the way, we’ve lost the habit, and the skill, of designing data properly. Not heavyweight diagrams and long documents, but simple, deliberate alignment across business, data and tech. This isn’t about slowing things down. It’s about making space for practical, agile data design at the point it matters most. 🧩 Define what data is needed alongside the process 📋 Align business analysts, product owners and data teams early ✔️ Build structure and validation into workflows from the start 👉 Because data should always be aligned to the business, and earn its place on the pitch. When that happens, everything else flows better. Curious how others are approaching this; are you bringing data into the room early enough?

98

28 Comments

I had the privilege of speaking at the Skills Development Network Digital Futures Midlands Conference this morning, sharing insights from our work on the NHS Federated Data Platform (FDP). It was energising to see so much engagement from colleagues across the region, from thoughtful questions to important conversations on data security, governance, and public trust. A few reflections that seemed to resonate in the room: 🔹 Delivering digital at scale is often less about shiny tech and more about aligning systems, people, and processes. 🔹 Integration across multiple source systems brings inevitable challenges, but also huge opportunities if tackled transparently and collaboratively. My main message was that we cannot afford to ignore delivery lessons, we must start as we mean to go on, looking internally for the subject matter expertise that will drive success. Sharing openly, what worked, what didn’t, and how we adapted, is how we build credibility, resilience, and trust in NHS digital. A big thank you to the organisers for bringing such a fantastic community together, and to my fellow panel members Asma Nafees Craig Kynaston for their candour. Always happy to connect with others facing similar challenges, the more we share, the stronger our collective digital future will be. #NHS #DigitalHealth #DataDriven #Leadership #DigitalFuturesMidlands

66

13 Comments

Who are the children living in poverty in Scotland? It's a simple question with a complex answer—and one that Spencer Thompson, Economic Advisor at the Scottish Government, set out to tackle through the Data Science and Innovation Accelerator. In his new blog, Spencer shares how he used machine learning to identify distinct clusters of households in poverty - offering policymakers a clearer, data-driven view of who needs support most. 📖 Read Spencer’s story https://lnkd.in/eageiXmm ✅ Today is the closing date for project applications to join the 2025 Data Science and Innovation Accelerator (if you're fast!) https://lnkd.in/egAPEBKk ✅ Or sign up as a mentor to help public sector staff tackle important challenges with data: https://lnkd.in/eT2YWMfG Chenglei Hu Tom Wilkinson #DataScience #ChildPoverty #ScottishGovernment #Innovation #PublicSectorInnovation #MachineLearning #SocialImpact #DSIA

31

3 Comments