Inside Reuters’ agentic AI video experiment

By Sara Guaglione * November 7, 2025 *

Ivy Liu

Reuters Pioneers AI-Driven Video Production to Accelerate Workflow

Reuters is advancing its use of artificial intelligence by integrating an AI-powered video producer to streamline and expedite video content creation. This marks a significant evolution from their traditional AI applications, which have primarily focused on enhancing text-based editorial processes.

From Text to Video: Expanding AI’s Role in Newsrooms

Historically, Reuters has leveraged AI to optimize text workflows, but Rob Lang, the newsroom AI editor, recognized the potential for agentic AI to revolutionize video editing. By employing AI to generate video metadata and automate the assembly of various edits, Reuters aims to reduce the manual labor involved in video production.

AI as a Collaborative Editor: Enhancing Efficiency Without Replacing Humans

Lang explained at the Digiday Publishing Summit Europe on October 28 that the AI system is currently capable of selecting key moments and compiling preliminary video cuts with impressive accuracy-even without multimodal capabilities. The AI’s role is to produce rough cuts, while final editing decisions remain under human supervision, exemplified by Reuters’ AI video producer, Enrique Flores Roldan.

Addressing Continuity and Contextual Challenges in Automated Editing

The AI’s assistance is particularly valuable for routine, time-intensive editing tasks. Lang envisions a future where the AI can intelligently manage continuity issues-such as changes in lighting caused by shifting sunlight or a subject removing their glasses-ensuring seamless video flow without human intervention.

Widespread AI Adoption Across Reuters’ Newsroom

Currently, approximately 60% of Reuters’ 2,500 journalists utilize AI tools, with 50 to 100 considered highly proficient users. These tools support a range of functions, from “vibe-coding” to investigative journalism assistance. Lang anticipates this adoption rate will climb by about 5% monthly, aiming for near-total integration by year-end, potentially reaching 80% usage.

Overcoming Resistance and Broadening AI Utility

One of the main hurdles to full AI adoption is the reluctance of some newsroom staff who do not yet see the value in AI. Lang shared an anecdote about a photographer unsure how AI could assist them, suggesting practical applications like using AI to find parking in unfamiliar cities. This highlights the broader potential of AI beyond journalistic tasks.

Enhancing Fact-Checking with AI and Proprietary Databases

Fact-checking remains a challenging area for AI due to difficulties in distinguishing credible information from unreliable sources. To address this, Reuters is developing a Retrieval-Augmented Generation (RAG) system-a specialized archive of Reuters’ own content that an LLM can query to produce more accurate and trustworthy outputs.

Lang emphasized that this RAG system provides a solid factual foundation for AI-assisted fact verification, improving the reliability of AI-generated content. Early experiments include integrating Reuters’ style guide into the database, enabling journalists to query specifics such as the correct spelling of Ukrainian President Volodymyr Zelensky’s name, which Reuters standardizes as “Zelenskiy.”

Innovative AI Tools Transforming Newsroom Productivity

  • Prompt Sandbox: A platform where newsroom staff input prompts and receive a series of AI-generated suggestions to aid content production.
  • Content Management System (CMS) Enhancements: Features like bullet point generators and headline creators streamline article development.
  • Automated Drafting: A CMS tool that synthesizes interview transcripts and press releases to produce initial article drafts.
  • Fact Genie: An AI-powered summarization tool that analyzes thousands of press releases weekly, recommending relevant stories for coverage.
  • Federal Bot: Developed by a Reuters journalist, this tool scans U.S. government data and press releases thrice daily, alerting reporters to noteworthy updates via targeted emails.

Looking Ahead: The Future of AI in Journalism

Lang reflects on the ongoing journey of integrating AI into newsroom operations, constantly exploring how these technologies will evolve and reshape journalism. Reuters’ commitment to innovation is evident in its continuous development of AI tools designed to enhance accuracy, efficiency, and editorial quality.

More in Media

More from this stream

Recomended