The AI Disclosures Project

Is Copyright Dead? (Weekly Roundup) What the latest Copyright Office report means for AI's training & inference, OpenAI's new head of monetization, Steve Bannon on AI regulations (bonus), and more.

New weekly roundup going in-depth on copyright after the head of the U.S. Copyright Office was seemingly fired for arguing not all AI training is protected under fair use.

This + model developers’ moves towards monetization, kooky comments from Steve Bannon & more!
open.substack.com/pub/asimovad...

Is the Internet's Business Model Finally Bust? (Weekly Roundup) Weekly Roundup 7 May, 2025: Matthew Prince makes a splash, OpenAI juices its benchmarks, 4o sycophant shows the need for post-deployment monitoring of behavioral risks, and more.

We’re back with another weekly roundup! This one covers some really shocking statistics from Cloudflare, typical OpenAI shenanigans, and more

open.substack.com/pub/asimovad...

Disclosures. I do not think that word means what you think it means. Disclosures are the language of networks and of markets

Great long read today from our co-director @timoreilly.bsky.social! He goes in depth about what we mean when we talk about disclosures and how they could shape safe AI development (hint: they’re not just a notice that something was AI generated)
open.substack.com/pub/asimovad...

Grok 3 on Elon Musk So you said you wanted the truth...

We wondered if Elon Musk’s control over Grok made it a tool for spreading misinformation. So we asked Grok itself to comment. It turns out being “maximally truth-seeking” doesn’t make it a huge fan of Musk…

open.substack.com/pub/asimovad...

Underappreciated Lessons from DeepSeek Most AI regulation misses the mark because regulators don't consider the real world context in which AI is deployed

Double feature today!

Tim O’Reilly goes long on what a people are missing from the DeepSeek mania of this week, specifically their (alleged) distillation of OpenAI’s models and what this tells us about cloud providers, terms of service, and disclosures

open.substack.com/pub/asimovad...

Weekly Roundup (January 29, 2025) An argument for open source, Microsoft's red team "goes rogue", covert benchmark funding, and inference-time compute for safety

New weekly roundup! Some thoughts on R1 (of course) and open source, along with new inference-time compute benefits, benchmark meddling, and more.

open.substack.com/pub/asimovad...

Weekly Roundup 1/6/2025 It's been way too long, so this is long too ;-)

New weekly roundup including more thoughts on how infrastructure will determine the future of AI, a review of the staggering AI development in 2024, and a great one-liner Sam Altman tweet
open.substack.com/pub/asimovad...

Born on the Fourth of July? The Bipartisan House AI Task Force Report Standards, industrial policy, and the very visible shadow of China

New blog post on the latest US AI Taskforce Report!

Co-director Ilan Strauss explores how it emphasizes the importance of international standard setting but misses the mark with regard to national security.

open.substack.com/pub/asimovad...

Clio's Blind Spot Why Clio ought to be reporting on API and business usage, not just individual users

New Substack post about Anthropic’s newly unveiled analytics tool for Claude—and why it has a huge blind spot:

open.substack.com/pub/asimovad...

The Architecture of Participation: Tim O'Reilly and Kevin Werbach on AI, Regulation, and Disclosure Podcast Episode. "Tim O'Reilly: The value of AI disclosure" in The Road to Accountable AI.

Our latest Substack post excerpts an interview of our cofounder on the architecture of participation. It’s a great conversation and covers a lot of what we’re researching. Read it here: open.substack.com/pub/asimovad...

Welcome to the AI Disclosures Project Bluesky! Through research, policy engagement, & industry collaborations we aim to learn how AI products are managed at the corporate level and how to disclose this. Knowledge is power.