Tech Policy Press

On the Tech Policy Press podcast, Justin Hendrix spoke to Mozilla CTO Raffi Krikorian about his "owners not renters" thesis: that the central AI question isn't open vs. closed, but whether we own AI or lease it from a handful of companies with their own incentives. Plus: his experience with Mythos. https://lnkd.in/eTZCYtXH

Trump’s new AI executive order establishes a “voluntary” framework for federal oversight of frontier AI models. Benjamin Lennett takes a close look at what’s in it and what’s next for the policy discussion. https://lnkd.in/eBJYnsJB

Global debates on AI tend to cast Africa as a landscape of opportunity and risk—but rarely as a site of normative authority, writes Leah Davina Junck, PhD from the Global Center on AI Governance. That should change, she says. https://lnkd.in/evHW9Srz

In its proposals, Brussels is walking a fine line between economic nationalism — now espoused by the world’s two economic superpowers — and the bloc’s traditional position of embracing free trade and market competition, writes Mark Scott. https://lnkd.in/eCEqjjbF

AI raises real concerns about jobs, creativity, education, and human agency, writes Shawn Murinko. But it is also removing barriers to access and enabling independence for people with disabilities. The challenge is ensuring AI is shaped with intention, care, and a commitment to equity. https://lnkd.in/eeWVFzPj

In my latest for Tech Policy Press on the UK CMA decision on Google, I put it in a broader context given developments in Brazil and Europe and caution that the extended timeframe for compliance and implementation is too slow.

Tech Policy Press has a new look! We've come a long way since we launched in 2021, and this new version of our website reflects our current editorial proposition, spotlights important work, and better serves our readers. We invite you to explore the updated site at techpolicy.press

If AI systems derive value from news, the organizations that invest in reporting, fact-finding, and accountability journalism must have leverage in determining the terms of that relationship. The UK CMA’s binding conduct requirement on Google is a first step toward that, writes Courtney Radsch, PhD. https://lnkd.in/e8t4-pF7

Germany's DeepL builds on AWS. France's Mistral builds with Nvidia. Sweden's Lovable just got a jump scare as Anthropic moved in to eat its lunch. This is what "European AI sovereignty looks like in spring 2026. It's also why Leevi Saari and I are launching a three-part series with Tech Policy Press and the AI Now Institute. This week's Sovereignty Package rests on a whole range of assumptions it never examines, including that more European AI companies automatically means a more sovereign Europe. Because the question that's more important than sovereignty in the abstract isn't how many companies Europe has. It's where the value they create actually ends up. Even when Europe builds its own AI, that value often flows elsewhere. That's because the reality is messy! While Europe still thinks in markets, the world thinks in ecosystems. What looks like a European AI ecosystem is a set of incentives pulling in different directions: exit horizons that don't reward independence, market structures that funnel revenue upstream to US providers, infrastructure dependencies that deepen with every round of investment. Most EU tech policy commentary works top-down - it analyses policy proposals. This series works bottom-up: who actually supplies whom, where revenue flows, which dependencies are already baked in. Three parts: ◼️ The application layer: how startups like Lovable become distribution channels for US models, and what Europe actually captures. ◼️ Infrastructure: why even Europe's strongest independents, like DeepL, end up dependent on US hyperscalers. ◼️ The synthesis: the shared structural dynamic, the strategic choices Europe keeps avoiding, and the honest uncertainties about where this is heading. What Europe could do differently is the harder question - and it's where the series lands. You can follow the series on Tech Policy Press or via the EU AI industrial policy newsletter https://lnkd.in/d_GxGKt7

States are rolling out generative AI tools to thousands of public employees, but governance and transparency standards often lag behind deployment, writes Javaid Iqbal Sofi. He discusses the sharply different approaches in New York, Pennsylvania, New Jersey, Colorado, and California. https://lnkd.in/e43pJNyi