Sign up to save your podcasts
Or
Share “The EU Is Asking for Feedback on Frontier AI Regulation (Open to Global Experts)—This Post Breaks Down What’s at Stake for AI Safety” by Katalina Hernandez
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
“The EU Is Asking for Feedback on Frontier AI Regulation (Open to Global Experts)—This Post Breaks Down What’s at Stake for AI Safety” by Katalina Hernandez
The European AI Office is currently writing the rules for how general-purpose AI (GPAI) models will be governed under the EU AI Act.
The are explicitly asking for feedback on how to interpret and operationalize key obligations under the AI Act.
This includes the thresholds for systemic risk, the definition of GPAI, how to estimate training compute, and when downstream fine-tuners become legally responsible.
Why this matters for AI Safety:
The largest labs (OpenAI, Anthropic, Google DeepMind) have already expressed willigness to sign on to the Codes of Practice voluntarily.
These codes will become the de facto compliance baseline, and potentially a global reference point.
So far, AI safety perspectives are severely underrepresented.
Input is urgently needed to ensure the guidelines reflect concerns around misalignment, loss of control, emergent capabilities, robust model evaluation, and the need for interpretability audits.
Key intervention points include [...]
---
Outline:
(00:43) Why this matters for AI Safety:
(02:05) Purpose of this post:
(03:04) TL;DR
(03:08) What the GPAI Codes of Practice will actually Regulate
(04:49) What AI safety researchers should weigh in on:
(05:39) 1. Content of the Guidelines
(06:22) 2. What counts as a General-Purpose AI Model ?
(07:13) 2.1 Conditions for Sufficient Generality and Capabilities
(09:23) 2.2 Differentiation Between Distinct Models and Model Versions
(10:30) Why this matters for AI Safety:
(11:35) 3. What counts as a Provider of a General-Purpose AI Model ?
(12:38) 3.1 What Triggers Provider Status?
(13:13) Downstream Modifiers as Providers
(14:08) 3.2 GPAI with Systemic Risk: Stricter Thresholds
(15:14) 4. Exemptions for Open-Source models
(16:24) 5. Estimating Compute: The First Scalable Safety Trigger
(17:22) 5.1 How to Estimate Compute
(18:13) 5.2 What Counts Toward Cumulative Compute
(19:16) 6. Other Legal & Enforcement details
---
First published:
April 22nd, 2025
Source:
https://www.lesswrong.com/posts/AbnfsnEEmHFmprGzm/the-eu-is-asking-for-feedback-on-frontier-ai-regulation-open
---
Narrated by TYPE III AUDIO.
View all episodes
By LessWrongThe European AI Office is currently writing the rules for how general-purpose AI (GPAI) models will be governed under the EU AI Act.
The are explicitly asking for feedback on how to interpret and operationalize key obligations under the AI Act.
This includes the thresholds for systemic risk, the definition of GPAI, how to estimate training compute, and when downstream fine-tuners become legally responsible.
Why this matters for AI Safety:
The largest labs (OpenAI, Anthropic, Google DeepMind) have already expressed willigness to sign on to the Codes of Practice voluntarily.
These codes will become the de facto compliance baseline, and potentially a global reference point.
So far, AI safety perspectives are severely underrepresented.
Input is urgently needed to ensure the guidelines reflect concerns around misalignment, loss of control, emergent capabilities, robust model evaluation, and the need for interpretability audits.
Key intervention points include [...]
---
Outline:
(00:43) Why this matters for AI Safety:
(02:05) Purpose of this post:
(03:04) TL;DR
(03:08) What the GPAI Codes of Practice will actually Regulate
(04:49) What AI safety researchers should weigh in on:
(05:39) 1. Content of the Guidelines
(06:22) 2. What counts as a General-Purpose AI Model ?
(07:13) 2.1 Conditions for Sufficient Generality and Capabilities
(09:23) 2.2 Differentiation Between Distinct Models and Model Versions
(10:30) Why this matters for AI Safety:
(11:35) 3. What counts as a Provider of a General-Purpose AI Model ?
(12:38) 3.1 What Triggers Provider Status?
(13:13) Downstream Modifiers as Providers
(14:08) 3.2 GPAI with Systemic Risk: Stricter Thresholds
(15:14) 4. Exemptions for Open-Source models
(16:24) 5. Estimating Compute: The First Scalable Safety Trigger
(17:22) 5.1 How to Estimate Compute
(18:13) 5.2 What Counts Toward Cumulative Compute
(19:16) 6. Other Legal & Enforcement details
---
First published:
April 22nd, 2025
Source:
https://www.lesswrong.com/posts/AbnfsnEEmHFmprGzm/the-eu-is-asking-for-feedback-on-frontier-ai-regulation-open
---
Narrated by TYPE III AUDIO.
More shows like LessWrong (30+ Karma)
View all
Making Sense with Sam Harris
26,369 Listeners
Conversations with Tyler
2,397 Listeners
The Peter Attia Drive
7,779 Listeners
Sean Carroll's Mindscape: Science, Society, Philosophy, Culture, Arts, and Ideas
4,103 Listeners
ManifoldOne
87 Listeners
Your Undivided Attention
1,451 Listeners
All-In with Chamath, Jason, Sacks & Friedberg
8,778 Listeners
Machine Learning Street Talk (MLST)
89 Listeners
Dwarkesh Podcast
355 Listeners
Hard Fork
5,370 Listeners
The Ezra Klein Show
15,053 Listeners
Moonshots with Peter Diamandis
460 Listeners
No Priors: Artificial Intelligence | Technology | Startups
126 Listeners
Latent Space: The AI Engineer Podcast
64 Listeners
BG2Pod with Brad Gerstner and Bill Gurley
432 Listeners