/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

days · browse · Enter similar · o open

An interview with OpenAI CTO Mira Murati on Sora, set for release later in 2024, its training data, red-teaming Sora, working with the film industry, and more

In an exclusive interview with our columnist, CTO Mira Murati explains the company's new Sora AI video tool and how it plans to roll it out

Wall Street Journal Joanna Stern

Context & Ripple Effects

Sora arrived amid a split reaction: its capabilities drew attention, while questions about the model’s training data remained unresolved. Murati’s account places safety testing, release planning and film-industry engagement at the center of OpenAI’s next step.

This matters because Sora’s path to users depends on more than output quality. The subsequent use of selected creative partners to showcase Sora illustrates how tightly managed access can also serve as product validation and industry outreach.

First-order effects

  • OpenAI’s near-term task is to gate Sora’s release through red-teaming and rollout decisions rather than make the video model broadly available at once.
  • Film-industry participants gain a direct channel into Sora’s development and use cases, while questions around its training data remain material to that engagement.

Second-order effects

  • Studios, agencies and creators evaluating generative video will have reason to seek clearer terms on training-data provenance, permitted uses and safeguards before integrating the tool into production workflows.
  • A selective rollout raises pressure on competing text-to-video developers to pair model demonstrations with credible access, safety and creator-partnership strategies.

Third-order effects

  • If controlled releases and industry partnerships become the standard route for frontier video models, deployment governance—not just model capability—will increasingly shape who can use these systems and in which workflows.
  • The unresolved training-data issue suggests that adoption in media may remain conditioned by rights and accountability arrangements, though the coverage does not establish what form those arrangements will take.

The trend: Generative-video AI is moving from attention-grabbing demonstrations toward governed deployment, with safety testing and creative-industry integration becoming part of product competition.