Google DeepMind and A24 Announce First-of-its-Kind Research Partnership

Title: AI Industry Confronts Partnerships, Legal Battles, Restrictions

01Google DeepMind and A24 Announce First-of-its-Kind Research Partnership

Google DeepMind and independent film studio A24 have announced a first-of-its-kind research partnership that will explore how frontier AI can be developed for media applications. The collaboration marks a novel approach to AI research, moving beyond traditional academic or enterprise partnerships to engage directly with creative industries. Unlike typical AI development deals, this partnership appears focused on fundamental research questions rather than immediate product deployment.

The collaboration signals a potential shift in how AI labs approach creative applications of their technology. Rather than licensing models to studios after development, DeepMind is bringing filmmakers into the research process itself. This could influence how future AI systems are designed with media production in mind. Industry observers see the partnership as an attempt to shape AI policy discussions from within the creative community, rather than facing external regulation.

02Midjourney Seeks to Compel Hollywood Studios to Disclose Their Own AI Usage

In an escalating legal dispute, Midjourney is seeking to compel three Hollywood studios to reveal how they themselves use AI in their production pipelines. The image generation company faces copyright infringement claims from the studios, but is now turning the tables by demanding transparency about the studios' own AI practices. Midjourney argues the studios are inconsistent in their stance on AI rights, using the technology while simultaneously suing AI companies.

The legal maneuvering highlights what critics describe as Hollywood's contradictory position on AI. Studios have publicly condemned AI training on copyrighted material while reportedly incorporating AI tools into their own workflows. Midjourney's legal team is positioning the discovery request as essential to understanding whether the studios genuinely object to AI usage or merely want to control who benefits from it. A ruling in Midjourney's favor could force studios to disclose the extent of their AI integration.

03Alibaba Reportedly Bans Employees from Using Claude Code

China's largest e-commerce company has reportedly prohibited employees from using Claude Code, classifying Anthropic's coding assistant as high-risk software. The ban prevents staff from accessing the AI-powered development tool through company systems, citing concerns about data security and potential intellectual property exposure. Alibaba reportedly views Claude Code as incompatible with internal security requirements.

The restriction reflects growing geopolitical tensions in the AI tooling space between US-based AI companies and Chinese enterprises. Anthropic's Claude models are developed in the United States and subject to American export regulations, which may complicate Alibaba's assessment of data handling risks. The ban positions Alibaba alongside other Chinese firms that have restricted foreign AI tools, potentially strengthening domestic alternatives. Analysts note this could accelerate development of Chinese-owned AI coding assistants as enterprises seek compliant alternatives.

Also today

04
Better Models: Worse Tools — Claude Opus 4.8 Invents Invalid Fields in Tool Calls — Newer Claude models including Opus 4.8 are generating malformed tool calls with invented fields that don't match schemas, causing edit tools to reject valid operations.
05
Open Source AI Gap Map Indexes 421 Products Across 228 Organizations — The Global Partnership for AI's new Gap Map catalogs 421 open source AI products including 266 tools, 85 models, and 50 datasets as a public option for AI development.
06
Google DeepMind Unionization Talks Off to Rocky Start — Employees at Google DeepMind expressed frustration during negotiations, saying executives are unwilling to meaningfully engage with unionization prospects.
07
Mark Zuckerberg Tells Staff AI Agents Haven't Progressed as Quickly as Hoped — Meta's CEO reportedly told employees at an internal meeting that AI development efforts are not moving as quickly as the company anticipated.
08
What is Mistral AI? Everything to Know About the OpenAI Competitor — French AI startup Mistral AI, which offers open source models and has raised significant funding, aims to put frontier AI in the hands of everyone.
09
New Google Commercial Imagines Declaration of Independence Written with AI — Google released a new commercial imagining what the Founding Fathers could have accomplished with Google Workspace AI tools, 250 years later.
10
June 2026 Newsletter: Claude Fable 5, GPT-5.6, GLM-5.2, and US Export Restrictions — Simon Willison's monthly newsletter covers major model releases including Claude Fable 5 and GPT-5.6, GLM-5.2 as the new best open weights model, and US AI export restrictions.
11
The Great Loops Debate and State of AI Engineering at AI Engineer World's Fair — The AI Engineer World's Fair concluded with debates about loops, a comprehensive state of AI engineering report, and keynotes focused on future development priorities.