🔍 In this TL;DR episode, Joan explains how Article 50 of the EU AI Act sets out high-level transparency obligations for AI developers and deployers—requiring users to be informed when they interact with AI or access AI-generated content—while noting that excessive labeling can itself be misleading. She highlights why the forthcoming Code of Practice must focus on clear principles rather than fixed technical solutions, ensuring transparency helps prevent deception without creating confusion in a rapidly evolving technological environment.

📌 TL;DR Highlights

⏲️[00:00] Intro

⏲️[00:33] Q1-What’s the core purpose of Article 50, and why is this 10-month drafting window so critical for the industry?

⏲️[02:31] Q2-What’s the difference between disclosing a chatbot and technically marking AI-generated media?

⏲️[06:27] Q3-What is the inherent danger of "too much transparency" or over-labeling content? How do we prevent the "liar's dividend" and "label fatigue" while still fighting deception?

⏲️[10:00] Q4-If drafters should avoid one rigid technical fix, what’s your top advice for building flexibility into the Code of Practice?

⏲️[13:11] Q5-Did you consult other stakeholders when developing your whitepaper analysis?

⏲️[16:45] Wrap-up & Outro

💭 Q1 - What’s the core purpose of Article 50, and why is this 10-month drafting window so critical for the industry?

🗣️ “Article 50 sets only broad transparency rules—so a strong Code of Practice is essential.”

💭 Q2 - What’s the difference between disclosing a chatbot and technically marking AI-generated media?

🗣️ “If there’s a risk of confusion, users must be clearly told they’re interacting with AI.”

💭 Q3 - What is the inherent danger of "too much transparency" or over-labeling content? How do we prevent the "liar's dividend" and "label fatigue" while still fighting deception?

🗣️ “Too much transparency can mislead just as much as too little.”

💭 Q4 - If drafters should avoid one rigid technical fix, what’s your top advice for building flexibility into the Code of Practice?

🗣️ “We should focus on principles, not chase technical solutions that will be outdated in months.”

💭 Q5 - What is the one core idea you want policymakers to take away from your research?

🗣️ “Transparency raises legal, technical, psychological, and even philosophical questions—information alone doesn’t guarantee real agency."

📌 About Our Guests

🎙️ Joan Barata | Faculdade de Direito - Católica no Porto

🌐 linkedin.com/in/joan-barata-a649876

Joan Barata works on freedom of expression, media regulation, and intermediary liability issues. He is a Visiting professor at Faculdade de Direito - Católica no Porto and Senior Legal Fellow at The Future Free Speech project at Vanderbilt University. He is also a Fellow of the Program on Platform Regulation at the Stanford Cyber Policy Center.

#AI #artificialintelligence #generativeAI

🔍 In this episode, Caroline and Alene unravel why the popular idea of “AI memorisation” leads policymakers down the wrong path—and how this metaphor obscures what actually happens inside large language models. Moving from the technical realities of parameter optimisation to the policy dangers of doctrinal drift, they explore how misleading language can distort copyright debates, inflate compliance burdens, and threaten Europe’s research and innovation ecosystem.

📌 TL;DR Highlights

⏲️[00:00] Intro

⏲️[00:33] Q1-In your view, what is the biggest misunderstanding behind the ‘memorisation’ metaphor in AI, and why is this framing so problematic when applied to copyright law?

⏲️[02:00] Q2-what actually happens inside a large language model during training, and explain why this process should not be treated as copyright ‘reproduction’?

⏲️[03:32] Q3-What do you see as the main legal, economic, and innovation risks for Europe if policymakers continue relying on the memorisation metaphor when designing AI regulation?

⏲️[04:39] Q4-If ‘memorisation’ is the wrong frame, what alternative concepts or policy focus areas should policymakers adopt to regulate AI more accurately and effectively?

⏲️[06:28] Q5-What is the one core idea you want policymakers to take away from your research?

⏲️[07:32] Wrap-up & Outro

💭 Q1 - In your view, what is the biggest misunderstanding behind the ‘memorisation’ metaphor in AI, and why is this framing so problematic when applied to copyright law?

🗣️ “A large language model is not a filing cabinet full of copyrighted material.”

💭 Q2 - what actually happens inside a large language model during training, and explain why this process should not be treated as copyright ‘reproduction’?

🗣️ "Training is parameter optimisation, not the storage of protected expression.”

💭 Q3 - What do you see as the main legal, economic, and innovation risks for Europe if policymakers continue relying on the memorisation metaphor when designing AI regulation?

🗣️ "Stretching the reproduction right to cover statistical learning would be disastrous for research and innovation in Europe.”

💭 Q4 - If ‘memorisation’ is the wrong frame, what alternative concepts or policy focus areas should policymakers adopt to regulate AI more accurately and effectively?

🗣️ "We need mechanism-aware regulation, not metaphor-driven lawmaking.”

💭 Q5 - What is the one core idea you want policymakers to take away from your research?

🗣️ "Don’t write rules for filing cabinets when we are dealing with statistical models.”

📌 About Our Guests

🎙️ Aline Larroyed | Dublin City University

🌐 linkedin.com/in/aline-l-624a3655

🌐 Article | The Fallacy Of The File: How The Memorisation Metaphor Misguides Copyright Law And Stifles AI Innovation

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5782882

Aline Larroyed is a postdoctoral researcher at Dublin City University and holds a PhD in International Law with a background in linguistics. She brings 20 years of experience in human rights, intellectual property, and international regulation, and is a member of the Institute for Globalization and International Regulation at Maastricht University and the COST LITHME network.

#AI #ArtificialIntelligence #GenerativeAI

🔍 In this TL;DR episode, Emmie Hine (Yale Digital Ethics Center) makes the case for Europe’s leadership in open-source AI—thanks to strong infrastructure, multilingual data, and regulatory clarity. With six key policy recommendations, the message is clear: trust and transparency can make EU models globally competitive.

📌 TL;DR Highlights

⏲️[00:00] Intro

⏲️[00:43] Q1-What advantages make the EU a strong contender in open-source AI compared to the US and China?

⏲️[03:08] Q2-How do EU regulations and initiatives enhance model trustworthiness and feasibility?

⏲️[05:43] Q3-What policy recommendations does the paper offer for EU AI governance and deployment?

⏲️[10:10] Wrap-up & Outro

💭 Q1 - What advantages make the EU a strong contender in open-source AI compared to the US and China?

🗣️ “The EU’s primary advantage is actually in regulatory leadership and trustworthiness.”

🗣️ “There’s access to a lot of multilingual data, which is really great and important for creating multilingual LLMs.”

🗣️ “As DeepSeek showed, big leaps are very possible.”

💭 Q2 - How do EU regulations and initiatives enhance model trustworthiness and feasibility?

🗣️ “Its framework of risk-based classification and how it encourages ethics by design—I think that’s really, really important for guiding responsible development and deployment of foundation models.”

🗣️ “The AI Office is really putting an emphasis on multi-stakeholder collaboration... bringing in civil society, academia, and industry.”

🗣️ “The Code of Practice will provide clarity on training data and copyright as well.”

💭 Q3 - Why focus on regulating specific AI apps instead of AI overall?

🗣️ “Establishing an EU-wide open-source foundation model governance framework... would combat open-washing, where companies say, ‘Oh yeah, this is open source,’ but it’s not really.”

🗣️ “A certification and benchmarking system to evaluate open models for security, reliability, ethics, and performance... would help boost user trust and also international competitiveness.”

🗣️ “Expanding funding programs—especially for SMEs and startups—can help them put out more competitive models while encouraging multilingual capabilities and ethical development.”

🗣️ “Ideally, this investment will ensure that these facilities are going to be energy efficient to help combat the climate impacts of models.”

🗣️ “There was definitely not a universal understanding of open-source technology in general, and specifically around open-source AI... So promoting digital literacy and responsible AI usage is going to be really, really crucial.”

📌 About Our Guest

🎙️ Emmie Hine | Yale Digital Ethics Center

🌐 Article | Open-Source Foundation Models Made in the EU: Why it is a Good Idea

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5191372

🌐Newsletter | The Ethical Reckoner

https://ethicalreckoner.substack.com/

🌐 Emmie Hine

linkedin.com/in/emmiehine

Emmie Hine is a Research Associate at the Yale Digital Ethics Center and a PhD candidate in Law, Science, and Technology at the University of Bologna and KU Leuven. Her research focuses on the ethics and governance of emerging technologies in different geopolitical contexts.

#AI #ArtificialIntelligence #GenerativeAI

🔍 In this TL;DR episode, Milton Mueller (the Georgia Institute of Technology School of Public Policy) argues that what we call “AI” is really just part of a broader digital ecosystem. Instead of vague, top-down AI regulation, he calls for context-specific rules that address actual uses—like facial recognition or medical diagnostics—rather than the underlying technology. The key message: regulate behavior and applications, not “AI” as an abstract concept.

📌 TL;DR Highlights

⏲️[00:00] Intro

⏲️[00:47] Q1-Why do you see AI as “distributed computing,” not a new tech?

⏲️[04:05] Q2-Why aren’t issues like misinformation and bias new to AI?

⏲️[09:29] Q3-Why focus on regulating specific AI apps instead of AI overall?

⏲️[17:02] Wrap-up & Outro

💭 Q1 - Why do you see AI as “distributed computing,” not a new tech?

🗣️ "Distributed computing is a revolutionary technology... it started around 1945, 1950, and it's been progressing and evolving ever since."

🗣️ "What people are calling AI is really a digital ecosystem, which is a cybernetic system that consists of computing devices, networks, data, and software."

🗣️ "When people talk about governing AI, it's not just about AI... you're going to have to control everything about ICT, everything about digital ICT."

🗣️ "If indeed you're talking about some way of controlling in advance the capabilities to do machine learning applications, you really are committing yourself to controlling every element of that digital ecosystem."

💭 Q2 - Why aren’t issues like misinformation and bias new to AI?

🗣️ "People say AI is a problem because of its capability to distribute misinformation, but if you go back to 1995, we were already noticing that some of the things you find on the internet are not true."

🗣️ "In the early days of the internet, people thought that just because something was online, it had to be credible or true."

🗣️ "What was giving you these search results... was an early form of artificial intelligence. You had robots searching the internet and scanning websites according to a program."

🗣️ "What’s interesting is that AI-driven search engines were already giving biased results back in the '90s... search engine bias was something people were complaining about."

🗣️ "You are interacting with an artificial intelligence mechanism that is distributed and globalized, and you're trying to control the results or objecting to the results in ways that are very similar to what we're hearing now about AI."

💭 Q3 - Why focus on regulating specific AI apps instead of AI overall?

🗣️ "The problems posed by a specific AI application would require different kinds of responses depending on the nature of the problem."

🗣️ "Can you solve a problem like that by regulating facial recognition technology? And the answer is no. What you do is you regulate police practice."

🗣️ "You have to be more specific. It's not about regulating AI in a general way; it's about regulating specific applications in their specific context."

🗣️ "Whenever you hear somebody say 'AI is causing this problem,' just substitute the word computing, and you'll see how it makes just as much sense."

🗣️ "The idea that you can regulate technology to prevent it from doing any harm in advance is an idea that, as a historian of technology, just doesn't make any sense."

📌 About Our Guest

🎙️ Milton Mueller | the Georgia Institute of Technology School of Public Policy

🌐 Article | It's just distributed computing: Rethinking AI governance

https://www.sciencedirect.com/science/article/pii/S030859612500014X

🌐 Milton Mueller

https://www.linkedin.com/in/miltonmueller/

Milton Mueller is Professor at the Georgia Institute of Technology School of Public Policy. He specialises in the political economy of information and communication and is the author of seven books and numerous journal articles. His work informs not only public policy but also science and technology studies, law, economics, communications, and international studies.

#AI #ArtificialIntelligence #GenerativeAI

🔍 In this TL;DR episode, Kevin Frazier (University of Texas at Austin school of Law) outlines a proposal to realign U.S. copyright law with its original goal of spreading knowledge. The discussion introduces three key reforms—an AI training presumption, research safe harbors, and data commons—to help innovators access data more easily. By reducing legal ambiguity, the proposals aim to support responsible AI development and level the playing field for startups and researchers.

📌 TL;DR Highlights

⏲️[00:00] Intro

⏲️[00:40] Q1-How do copyright laws limit AI’s training data and knowledge diffusion?

⏲️[04:05] Q2-How does the original intent of the Intellectual Property Clause conflict with current copyright rules?

⏲️[10:35] Q3-What reforms could align copyright with promoting science?

⏲️[15:36] Wrap-up & Outro

💭 Q1 - How do copyright laws limit AI’s training data and knowledge diffusion?

🗣️ "Copyright law was designed to promote knowledge creation. Now it functions as a bottleneck in that knowledge ecosystem."

🗣️ "Today's AI systems face a paradoxical constraint: they require vast oceans of human-created content to learn, and yet our copyright framework increasingly cordons off those essential waters."

🗣️ "Early publishers faced guild restrictions and royal monopolies that limited the dissemination of knowledge. Today's AI developers are navigating a similarly restrictive landscape through the barriers of copyright law."

🗣️ "When there's ambiguity in the law, that hinders innovation. It leads to litigation and poses a real threat to startups trying to determine whether they can use copyrighted information for training."

💭 Q2 - How does the original intent of the Intellectual Property Clause conflict with current copyright rules?

🗣️ "The IP clause starts off with a mandate that Congress spread knowledge. If something isn’t promoting the progress of science, then it can’t be interpreted as constitutional."

🗣️ "Copyright began as a 14-year term. Now it's expanded to more than 70 years — a huge restriction on the ability to spread knowledge."

🗣️ "The founders hated monopolies. They’d seen how royal prerogatives were used to quash innovation — and tried to create a better system for incentivizing knowledge."

🗣️ "AI tools are unparalleled in their ability to create knowledge. The question now is: can we spread that knowledge?"

🗣️ "The Constitution’s goal wasn’t just to reward creators — it was to spread science and useful arts as far and wide as possible."

💭 Q3 - What reforms could align copyright with promoting science?

🗣️ "We need a clear statutory presumption that using works for machine learning constitutes fair use — that sort of clarity is essential for startups and research institutions to compete."

🗣️ "Without robust datasets, the positive use cases of AI — from public health breakthroughs to AI tutors for differently-abled students — simply aren’t possible."

🗣️ "Imagine if all that data your smartwatch gathers went toward training AI models tailored to the public good — that’s the promise of data commons."

🗣️ "AI is like fire: it can spread and fuel incredible progress — but if we smother it with fire extinguishers too soon, only the biggest players will be able to benefit."

🗣️ "We must make sure this isn’t a world where only OpenAI, Anthropic, and Google build the models — we need a future with many options and many positive use cases of AI."

📌 About Our Guest

🎙️ Kevin Frazier | the University of Texas at Austin School of Law

🌐 Article | Progress Interrupted: The Constitutional Crisis in Copyright Law

https://jolt.law.harvard.edu/digest/progress-interrupted-the-constitutional-crisis-in-copyright-law

🌐 Kevin Frazier

https://www.linkedin.com/in/kevin-frazier-51811737/

Kevin Frazier is an AI Innovation and Law Fellow at the Austin School of Law of the University of Texas. He is also a Contributing Editor at Lawfare, a non-profit publication and he developed the first open-source Law and AI syllabus.

#AI #ArtificialIntelligence #GenerativeAI

🔍 In this TL;DR episode, Paul Keller (The Open Future Foundation) outlines a proposal for a common opt-out vocabulary to improve how EU copyright rules apply to AI training. The discussion introduces three clear use cases—TDM, AI training, and generative AI training—to help rights holders express their preferences more precisely. By standardizing terminology across the value chain, the proposal aims to bring legal clarity, promote interoperability, and support responsible AI development.

📌 TL;DR Highlights

⏲️[00:00] Intro

⏲️[00:41] Q1-Why is this vocabulary needed for AI training opt-outs?

⏲️[04:17] Q2-How does it help creators, AI developers, and policymakers and what are some of the concepts?

⏲️[11:55] Q3-What are its limitations, and how could it evolve?

⏲️[14:35] Wrap-up & Outro

💭 Q1 - Why is this vocabulary needed for AI training opt-outs?

🗣️ "At the core of the EU copyright framework is... the TDM exceptions – the exceptions for text and data mining that were introduced in the 2019 Copyright Directive."

🗣️ "It ensures that rights holders have some level of control over their works, and it makes sure that the majority of publicly available works are available to innovate on top of, to build new things."

🗣️ "The purpose of such a vocabulary is to provide a common language for expressing rights reservations and opt-outs that are understood in the same way along the entire value chain."

🗣️ "This vocabulary proposal is the outcome of discussions that we had with many stakeholders, including rights holders, AI companies, policymakers, academics, and public interest technologists."

💭 Q2 - How does it help creators, AI developers, and policymakers and what are some of the concepts?

🗣️ "At the very core, the idea of vocabulary is that you have some common understanding of language... that terms you use mean the same to other people that you deal with."

🗣️ "We offer these three use cases for people to target their opt-outs from... like sort of the Russian dolls: the wide TDM category that is AI training, and in that is generative AI training."

🗣️ "If all of these technologies sort of use the same definition of what they are opting out, it becomes interoperable and it becomes also relatively simple to understand on the rights holder side."

💭 Q3 - What are its limitations, and how could it evolve?

🗣️ "The biggest limitation is... we need to see if this lands in reality and stakeholders start working with this."

🗣️ "These information intermediaries... essentially convey the information from rights holders to model providers—then it has a chance to become something that structures this field."

🗣️ "It is designed as a sort of very simple, relatively flexible approach that makes it expandable."

📌 About Our Guest

🎙️ Paul Keller | The Open Future Foundation

🌐 Article | A Vocabulary for opting out of AI training and other forms of TDM

https://openfuture.eu/wp-content/uploads/2025/03/250307_Vocabulary_for_opting_out_of_AI_training_and_other_forms_of_TDM.pdf

🌐 Paul Keller

https://www.linkedin.com/in/paulkeller/

Paul Keller is the co-Founder and Director of Policy at the Open Future Foundation, a European nonprofit organization. He has extensive experience as a media activist, open policy advocate and systems architect striving to improve access to knowledge and culture.

#AI #ArtificialIntelligence #GenerativeAI