US GAO Framework for Assessing AI Competitiveness

What’s been most interesting while putting together this event is seeing just how many organisations are now asking the same question: 𝗛𝗼𝘄 𝗱𝗼 𝘄𝗲 𝗺𝗼𝘃𝗲 𝗯𝗲𝘆𝗼𝗻𝗱 𝗲𝘅𝗽𝗲𝗿𝗶𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻 𝗮𝗻𝗱 𝗰𝗿𝗲𝗮𝘁𝗲 𝗺𝗲𝗮𝘀𝘂𝗿𝗮𝗯𝗹𝗲 𝘃𝗮𝗹𝘂𝗲 𝘄𝗶𝘁𝗵 𝗔𝗜? The conversation has clearly shifted. It’s no longer just about trying tools or running pilots. The real challenge now is 𝗹𝗲𝗮𝗱𝗲𝗿𝘀𝗵𝗶𝗽 𝗮𝗹𝗶𝗴𝗻𝗺𝗲𝗻𝘁, 𝗼𝗿𝗴𝗮𝗻𝗶𝘀𝗮𝘁𝗶𝗼𝗻𝗮𝗹 𝗿𝗲𝗮𝗱𝗶𝗻𝗲𝘀𝘀, 𝗴𝗼𝘃𝗲𝗿𝗻𝗮𝗻𝗰𝗲, 𝘄𝗼𝗿𝗸𝗳𝗼𝗿𝗰𝗲 𝗮𝗱𝗼𝗽𝘁𝗶𝗼𝗻, and 𝗳𝗶𝗴𝘂𝗿𝗶𝗻𝗴 𝗼𝘂𝘁 𝘄𝗵𝗮𝘁 𝘀𝗰𝗮𝗹𝗶𝗻𝗴 𝗔𝗜 𝗮𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝗹𝗼𝗼𝗸𝘀 𝗹𝗶𝗸𝗲 inside established businesses. I am really looking forward to bringing together leaders across industries for an honest discussion around what’s working, what’s not, and where businesses are genuinely seeing impact. Very excited to see companies from across the Danish-UK Association's network already joining the conversation. Thank you to everyone who has signed up so far. If this is a topic close to your work, grab your ticket now! There are limited seats. 

According to The New York Times, DeepSeek AI’s open-source AI model has given China a significant soft power advantage through low cost and global accessibility. As the U.S. keeps its systems closed, the AI competition is becoming not only technological but also ideological. 🖇️ Katrin Bennhold

For our #GlobalAIWatch column, we’re highlighting an analysis from OECD.AI by Charles Martinet and Yohann Ralle that we found thought-provoking. The piece examines whether mid-sized economies face a binary choice between relying on US- or China-based frontier AI systems and pursuing cross-border cooperation as a credible alternative. Drawing on examples of shared compute, federated learning, and multinational research initiatives, the authors examine how advanced AI development can be pursued without sacrificing sovereignty or control over data. We’re sharing this piece because it surfaces practical questions policymakers are grappling with: who builds frontier AI, who pays for it, and what preserves cooperation. 👉 #RebootDemocracy re-post: https://lnkd.in/eMw2EW9p 👉 Original OECD.AI piece: https://lnkd.in/eyZ7hk5E

For our #GlobalAIWatch column, we’re highlighting an analysis from OECD.AI by Charles Martinet and Yohann Ralle that we found thought-provoking. The piece examines whether mid-sized economies face a binary choice between relying on US- or China-based frontier AI systems and pursuing cross-border cooperation as a credible alternative. Drawing on examples of shared compute, federated learning, and multinational research initiatives, the authors examine how advanced AI development can be pursued without sacrificing sovereignty or control over data. We’re sharing this piece because it surfaces practical questions policymakers are grappling with: who builds frontier AI, who pays for it, and what preserves cooperation. 👉 #RebootDemocracy re-post: https://lnkd.in/eMw2EW9p 👉 Original OECD.AI piece: https://lnkd.in/eyZ7hk5E

How China Compete With US on AI When US Maintains Capabilities Edge ✅️China does not need to match the top US frontier model one-for-one to narrow the gap in real-world use. ✅️On OpenRouter, a platform where developers choose between competing AI models, Chinese models overtook US models in usage in February 2026 and now account for roughly four times the token volume. ✅️For many enterprise tasks, smaller open models can be good enough, especially where cost, customization or local deployment matter. Read more https://lnkd.in/gTn9E_9D Join Bloomberg analysts on May 13 for a conversation about US-China relations ahead of President Donald Trump’s planned visit to Beijing. Register at https://lnkd.in/gwgERiTm #AI #China #US

Kenny Shui, Vice President of 團結香港基金 Our Hong Kong Foundation (OHKF) and Executive Director of Public Policy Institute, Alex Mak, CFA, CESGA, Senior Research Manager of OHKF, Joyce Chan, Researcher of OHKF, and Hing Chak CHAN, Research Assistant of OHKF, argued that as the socio-economic impacts of artificial intelligence (AI) become increasingly evident, the cost of policy intervention rises day by day. Against the backdrop of rapidly evolving global AI technologies, how should Hong Kong utilize appropriate policy tools to avoid the dilemma of timing and missing development opportunities? As neighboring regions increasingly view AI as a "national necessity", how can Hong Kong promote the deep integration of AI with technological innovation and industrial development through systematic strategic coordination? In formulating this dedicated blueprint, how should the HKSAR Government uphold a people-centric principle and widely incorporate views from all stakeholders to ensure policies closely align with societal and industry needs, thereby creating maximum socio-economic value? Read More: https://bit.ly/4tJwxLw 2025 Technology and Innovation Report: Inclusive Artificial Intelligence for Development:https://bit.ly/4dg4crk #團結香港基金 #OurHongKongFoundation #OHKF #人工智能 #AI發展藍圖 #政策統籌 #技術和創新報告 #社會經濟價值 #ArtificialIntelligence #AIBlueprint #PolicyCoordination #TechnologyAndInnovationReport #SocioEconomicValue

The main idea here isn’t crazy. AI can erode agency when humans stop structuring choices, stop making decisions, or lose the institutional ability to reverse course. RAND Corporation frames this through “decisive coalitions,” coalition size, and whether AI starts shaping the menu of options itself. That part is useful. But this line: “Policymakers should consider minimum requirements for human presence in decisive coalitions.” It sounds like a safeguard but it isn’t. “Human presence” is not the same as human control. You can have a human in the loop who didn’t define the options, doesn’t understand the system, and isn’t positioned to challenge it. By the time they’re involved, the decision has already been made. They’re selecting from a pre-shaped menu. That’s not agency. The problem with framing it this way is that it becomes another compliance target. Organizations won’t optimize for real decision authority, they’ll optimize for meeting the requirement. The minimum human requirement. You get a human in the room, a signature on the output, and the appearance of oversight, while the AI system that shaped the outcome remains untouched. We’ve seen this before…where safeguards drift into liability shields. If this conversation is serious, the questions to ask are straightforward: who defines the options, who can override the system, what happens when they do, and who owns the consequences. Anything less is theater. Just more paperwork.

Our Head of Growth Luke Ashmore-Delaney has been ruminating on the limitations of AI and why we still need experts in the field. In business, answers are only as valuable as the questions that are asked. Check out his full piece in the comments:

Luke Ashmore-Delaney reckons the clock hasn't fully run out for human jobs, and I'm (only a little selfishly) inclined to agree - Check out his post, written for humans, by a human!

AI Minister Evan Solomon believes the government must build for the AI future. At BetaKit's #MostAmbitious: Town Hall, he said the government must be open to opportunities, but keep Canadian values in mind as it builds.