Alex Gluhak

We have a once-in-a-generation opportunity to build a sovereign UK AI Chip industry. The global market for AI chip products (accelerators, memories, fabrics etc) is growing at a staggering 30% per year -- and at this rate will be half the entire global semiconductor market by 2030. Just think of the economic growth potential if the UK could capture just 10% of the global market? We have deep AI talent in companies like DeepMind, OpenAI, ARM, and MSFT. And government commitment spawned by the AI Opportunities Plan and ARIA’s Scalable Compute program. What we need now is a BIG increase in the number of students studying electrical/electronic engineering and chip design.  Explore the detailed report here: [Building a Sovereign AI Chip Design Industry in the UK] https://lnkd.in/eTMxKN8c

329

30 Comments

The UK Ministry of Defence just published their first Responsible AI Senior Officers' Report, and the findings mirror what we're seeing across every sector: 𝐭𝐡𝐞 𝐝𝐞𝐦𝐚𝐧𝐝 𝐟𝐨𝐫 𝐀𝐈 𝐆𝐨𝐯𝐞𝐫𝐧𝐚𝐧𝐜𝐞 𝐏𝐫𝐨𝐬 𝐟𝐚𝐫 𝐨𝐮𝐭𝐬𝐭𝐫𝐢𝐩𝐬 𝐬𝐮𝐩𝐩𝐥𝐲. The MOD appointed Responsible AI Senior Officers (RAISOs) across 12 organisations. But the report makes clear these senior roles are just the start. Each RAISO needs dedicated support teams, local points of contact, and multidisciplinary expertise to function effectively. Their experience maps directly onto the four essential tracks for AI governance: ➜ Building organisational foundations: RAISOs are establishing governance structures, building AI inventories, and creating networks. The ad hoc approach worked initially, but now they need dedicated teams. ➜ Navigating laws and standards: JSP 936, their AI policy directive, has provided the impetus for internal policy updates across organisations. Good policy creates clarity that accelerates responsible adoption. ➜ Creating guardrails: They've embedded an AI ethics risk assessment toolkit into existing frameworks while facing what they describe as "the tri-fold task of recruiting external AI experts, training existing personnel to a high level, and upskilling the rest of the workforce." ➜ Operating with human oversight: They're building monitoring dashboards and launching Digital Skills for Defence, recognising that meaningful human control requires specific capabilities, training, and infrastructure. The report's conclusion on workforce is clear: "While there is a growing cadre of AI professionals in some areas, the most significant challenge identified across all RAISO statements is increasing the availability of suitably qualified and experienced personnel to support Defence's commitment to Responsible AI Adoption." What are the challenges and opportunities across these four tracks in your organisation? Download the skills map to see where your experience fits and plot your course to becoming an AI Governance Pro: 📥  https://lnkd.in/gHYeM4bM

200

9 Comments

Today, OpenAI and the British government signed a new agreement to harness AI’s transformative power to make government services more efficient, while fueling private-sector innovation across the UK. It’s another step forward in building a global system of democratic AI designed to put the technology into the hands of as many people as possible OpenAI CEO and co-founder Sam Altman signed the agreement remotely alongside Peter Kyle MP, the UK’s Secretary of State for Science, Innovation and Technology. As Sam put it, “AI is a core technology for nation building that will transform economies and deliver growth” He’s exactly right. As part of this partnership, we’ll work with the UK government to bring AI’s benefits into everyday use—helping civil servants work more efficiently, supporting small businesses as they navigate public services, and sharing technical expertise to advance safe, responsible deployment This partnership will advance the core goal of the UK’s AI Opportunities Action Plan, which calls for making AI more accessible by accelerating infrastructure construction and expanding access to the data needed to train cutting‑edge models AI is already having a tangible impact in Britain. Developers, startups, established companies like Virgin Atlantic, and world‑class research institutions such as Oxford University use our technology to fuel innovation, increase productivity, and create new economic opportunities. The UK government is leading by example: A GOV.UK business chatbot helps thousands of small enterprises get clearer, faster answers on rules and support programs “Humphrey,” Whitehall’s AI assistant, lightens administrative load so officials can focus on higher‑value work The agreement also reflects our ongoing commitment to working with the UK to bring AI’s gains to more British citizens. OpenAI’s first international office opened in London in 2023, and we plan to further expand our UK presence later this year Our work in the UK and around the world is focused on ensuring that democratic AI—which puts the technology in the hands of as many people as possible so they can share in its productivity gains—wins out over  concentrated AI, which would keep the benefits limited to a few This week’s expected release of the Trump Administration’s AI Action Plan is likely to add further momentum to the growing effort to build a global system of democratic AI rails by proposing new ways of speeding AI infrastructure construction, modernizing the grid, and generating the energy needed to power advanced AI data centers This is how you build democratic AI—not in one leap, but step by step, putting the technology in more hands so people everywhere can solve hard problems and improve their lives. This agreement is a powerful step forward Read more about the new partnership here https://lnkd.in/gUyUq2uB

375

11 Comments

Delighted to have contributed to the Java Sea Wall Report alongside Deloitte. Climate resilience is not simply about defense against water, it’s about securing economic stability, livelihoods, and long-term investment confidence. True resilience requires seeing infrastructure as both protection and opportunity, delivering security today while enabling growth tomorrow. This report is one step forward in rethinking how we finance and deliver resilience at scale. Read the report below.

50

4 Comments

The UK Government has just published its Compute Roadmap, outlining a long-term strategy to position the UK as a global leader in compute capability, spanning classical, AI, and quantum technologies. Of particular interest is the recognition that #quantumcomputing will become essential in addressing challenges that cannot be solved efficiently through classical means—such as molecular #modelling, secure #communications, and large-scale #optimisation. The roadmap also calls for the development of hybrid computing architectures, integrating quantum with #AI and #HPC, and stresses the need for application-specific quantum innovation in #chemistry, #materials science, and beyond. Importantly, the strategy places emphasis on building a coherent national compute #infrastructure and nurturing a highly skilled #workforce - both of which are critical for scaling the impact of quantum technologies. A significant signal of intent, and a valuable reference point for those working at the intersection of science, industry, and public policy. Link to Policy Paper: https://lnkd.in/eT9rTxGu Peter Kyle Chris Bryant Feryal Clark Baroness Jones of Whitchurch Emran Mian Joanna Davinson Lindsay Mason Chris Johnson Dave Smith Tom Scott Tom Newby Tianna V. Faiyaz Amin Rachel Maze David Knott Diego Rodriguez Mejias #QuantumComputing #UKScienceAndTechnology #ComputeStrategy #UKInnovation #AI #HPC #HybridComputing #Policy #DeepTech #ResearchAndDevelopment

61

Towards Competitive Cloud Ecosystems: Strategic Responses for Europe’s Digital Future | published by AI & Partners This report provides a structural assessment of #Europe’s cloud computing landscape, analysing how concentration in infrastructure and software markets undermines #digitalsovereignty and distorts competition. With five global providers currently controlling over 75% of Europe’s #cloud infrastructure market, and proprietary software licensing practices affecting more than 60% of EU-based enterprise clients, the report frames vendor lock-in and data portability restrictions as systemic barriers to #economicresilience and #digitalautonomy across the European Union. The study integrates legal analysis, expert interviews, and comparative policy benchmarking, revealing that only 9 EU Member States have implemented full interoperability mandates, while over 70% of public sector cloud contracts remain dependent on non-EU providers. Despite the entry into force of the Digital Markets Act (DMA) and Data Act, the report highlights ongoing circumvention tactics such as restrictive bundling, proprietary #APIs, and excessive egress fees. These practices increase customer switching costs by up to 35%, and in some sectors, limit multi-cloud adoption to below 20% of firms. The report outlines a three-pillar response strategy. First, it calls for revised antitrust enforcement focused on dynamic licensing abuses, supported by the creation of a dedicated Cloud Competition Unit under the European Commission. Second, it recommends the adoption of mandatory interoperability and portability protocols across #IaaS, #PaaS, and #SaaS layers, tied to certification mechanisms under the EU #Cybersecurity Act. Third, it encourages the allocation of at least 25% of public cloud procurement budgets toward certified #Europeanproviders to stimulate demand-side competition and reduce systemic dependency on non-EU actors. In conclusion, the report argues that Europe must reorient its cloud governance paradigm to favour anticipatory regulation, enforceable technical standards, and strategic industrial policy. As cloud infrastructure becomes a critical enabler of #AI, #quantumcomputing, and #crossborderdataecosystems, the ability to ensure open, contestable, and interoperable digital markets will define the EU’s technological sovereignty and long-term competitiveness.

38

1 Comment

The NHS 10-Year Plan Digital Ambition and the Fatal Mistake of Planning Forwards The newly published NHS 10-Year Plan should be a rallying cry for digital transformation. Instead, it reads like a manual for polishing the machinery of a crumbling paradigm. Despite the rhetoric, digital and AI are presented not as foundational capabilities, but as tools to marginally improve an already outdated system. This is not transformation. It is optimisation dressed up as strategy. The core issue is this. No one in the policymaking, management or delivery apparatus of the NHS appears to be asking the only question that matters: what kind of future health system do we actually need to build? And how do we get there? Instead of vision, we get bureaucracy. Instead of bold thinking, we get incrementalism. This is planning in its most reductive form — starting from the present, projecting forwards, and hoping that a few digital tweaks to EHRs and the NHS app will prepare us for the world that is coming. But the world is not waiting. The convergence of AI, systems biology, real-time data, and participatory technologies is redefining what is possible in health. The future is being built elsewhere, by people who understand that digital is not a department or a dashboard. It is the intelligence layer of the system itself. By contrast, the NHS plan treats digital as a side project. It is a bolt-on to the Horizon 1 machine, not the operating system of a reimagined future. Back in 2018, I led Beyond the Fog, a foresight report for the Royal Free Charity. We took a future-back approach and asked what a better, technology-enabled health system could look like. What we found was clear. If we are serious about lifelong health, we need to build a dynamic, adaptive, real-time, collaborative information infrastructure. One that serves not just patients, but clinicians, carers, researchers, and communities. One that enables learning, coordination, prevention and personalisation to happen in real time, everywhere. In that future, digital and AI are not add-ons. They are the brain and nervous system of the system. Without this core capability baked into the heart of a future NHS, there is no possibility of building a 21st-century health service. To outsource this strategic intelligence to 3rd party providers is to hand over the future of public health through the back door. It is not a clever partnership. It is a loss of sovereignty. With the launch of GPT-5 and the acceleration of biomedical science, this level of strategic blindness is indefensible. It is a dereliction of responsibility. A failure to lead. A failure to even try to understand the magnitude of the shift underway. The NHS does not need more cautious planning. It needs bold vision, serious strategic design, and the courage to build the capabilities that will define the next 50 years. If this plan is the best we can do, then we are not just behind the curve. We are facing the wrong direction.

149

55 Comments

The comment period for the proposed revised NEVI guidelines closed recently, and you can now view all 28 comments. I browsed 10 of the 28 and spoke with David Ferris at POLITICO this morning, primarily around what seems to be the most contentious issue ... ... FHWA encouraging states to “select charging locations where the charging station operator is the site host." As you might expect, almost all of the comments that included this issue were against it — except for the groups representing gas stations, convenience stores, and travel stops. While my comments didn't make the article, our data did - but my personal take is: Encouraging states to “select charging locations where the charging station operator is the site host" — sounds rationale (in theory less contract issues, host is also the operator and has a greater vested interest, can move more quickly, etc.). But ... It ignores the actual track record of reliability and customer experience. Just because a site host such as a convenience store chain is also the CPO/network doesn't guarantee great reliability, quicker response time, or a better customer experience. Tesla and IONNA, for example, both partner with Sheetz stores. There is no guarantee that Brand X convenience store across the street from one of these sites would be opened more quickly, more reliable, and provide a better experience — if it was also the owner/operator. Instead of encouraging a specific owner/site host model, FHWA should encourage stations that provide the best overall customer experience and awarding NEVI sites to applicants that have proven strong reliability, uptime, responsiveness, and a great customer experience. We at Paren are starting to see a lot of interest in this data by states which have been getting pressure from constituents and elected officials for awarding sites to applicants that aren't delivering on promises. What FHWA should encourage is greater accountability and that states award sites to the best applicants, not just those that fit a certain criteria. Secondly, what confounds me in all of this is that FHWA is merely"encouraging." What does that mean? States issue RFPs with scoring rubrics — so I guess this means in essence that FHWA is encouraging a row in the rubric that provides additional points to site hosts that are also the applicant/operator. If this happens, I can see applicants (that aren't also the site host) increase their match percentage to offset the points they might lose to competing applicants. I don't know the story behind this "guideline" - but it feels a bit fishy to me and is not consistent with "giving the states more flexibility." https://lnkd.in/gQ3mSgpj

61

15 Comments

Bringing together complex systems, network science, and computational social science yields powerful, vital insights on human-AI coevolution. This important paper by leaders including Albert-Laszlo Barabasi and Alex 'Sandy' Pentland describes the vital role of feedback in human-AI systems and the implications. Some of the key findings: 🔁 Feedback loops are the engine of coevolution. Human-AI coevolution is driven by a self-reinforcing feedback loop: users generate data that train AI, which in turn influences users’ future choices. This dynamic can amplify unintended consequences, such as polarization, bias, or system degradation. The loop is especially evident in recommender systems and generative AI models. 🌍 Society-centered AI reframes the debate. The authors introduce “society-centred AI” as a new paradigm, expanding beyond technology- or human-centred approaches. It emphasizes that the feedback loop's impact spans individual to societal levels, and managing it requires scientific innovation, legal regulation, and political action, not just more technology. 📉 Generative AI risks collapsing into sameness. In content generation, when LLMs are fine-tuned on AI-generated data, they regress to the mean—standardizing language and reducing diversity. This “autophagy” loop risks a slow erosion of linguistic and cultural variation. 📈 Recommenders reinforce inequality and monopolies. "Recommenders" - systems in retail, media, or navigation that guide users - increase efficiency for individuals but can reduce diversity and favor dominant players at scale. For instance, they may concentrate traffic on certain routes or drive sales toward popular items—reinforcing market concentration and inequality. 🏙️ AI can backfire on cities. Navigation apps like Google Maps aim to reduce individual travel time, but coevolution causes collective issues. Too many users being routed through “optimal” paths creates new congestion, longer travel times, and higher emissions. ⚖️ Ownership of recommenders shapes power. The authors highlight a modern twist on Marx’s theory: owning the “means of recommendation” (e.g., recommender platforms) confers immense influence. In the absence of governance, this could entrench digital monopolies, skew public discourse, and diminish collective welfare. 🚫 Techno-solutionism is insufficient. The paper cautions against the belief that more or better AI will fix AI’s harms. Without societal oversight and a shift from individual utility to collective benefit, the feedback loop will likely continue to deepen existing social divides, especially along lines of race, income, and geography.

47

14 Comments