This workshop will be held in english
A clear AI policy is an essential foundation for responsible AI use in a newsroom. However, for a policy to be effective, its principles must be translated into concrete rules, responsibilities, approval processes, and workflows that journalists and editors can apply in real situations.
A 2026 review by the Center for News, Technology & Innovation, drawing on 30 recent studies of AI governance in journalism, found that newsroom AI policies often emphasize important principles such as transparency, human oversight, and verification, but do not always provide sufficient guidance on how those principles should be implemented in daily editorial work.
This is where AI governance becomes essential. Governance connects the policy to practice. It determines who can approve an AI tool or workflow, which uses require additional oversight, how risks are assessed, who verifies AI-assisted work, when content should be disclosed or labelled, and what happens when an error or incident occurs.
"From AI Policy to AI Governance: Build Rules Your Newsroom Can Actually Use" is a hands-on workshop that will help participants answer these questions and turn responsible AI principles into practical newsroom decisions.
Participants will work on both parts of the challenge. They will begin developing or strengthening an AI policy, while also creating the governance elements needed to put that policy into practice.
The workshop is grounded in my work with newsrooms and media professionals on AI literacy, responsible adoption, editorial workflows, internal guidelines, and AI policy development. Through this work, I repeatedly encountered similar challenges: vague language, unclear responsibilities, inconsistent disclosure practices, missing rules for high-risk uses, insufficient attention to third-party tools and data protection, and policies that journalists could not easily apply in their daily work.
These recurring challenges led me to develop the free AImpactful AI Governance Builder, a guided tool and workshop, that helps newsrooms create or improve their public AI policies, internal guidelines, disclosure rules, risk registers, governance structures, incident-response procedures, and own training plans.
They will use its methodology during the workshop and will be able to continue developing their policies and governance documents afterwards, involve colleagues, and adapt the materials to the needs, size, and resources of their own organization.
The 90-minute workshop will include five practical stages:
1. Map current and planned AI use
Participants will identify where AI is already being used or considered across the newsroom.
This first step will help participants connect the policy to real workflows rather than treating AI governance as an abstract exercise.
2. Assess and classify risk
Participants will examine how the level of risk changes depending on the tool, task, content, data, and editorial context.
They will distinguish lower risk uses from sensitive or high-risk applications.
3. Build the core policy rules
Participants will begin defining the central rules of their AI policy, including:
• permitted, restricted, and prohibited uses;
• requirements for human review and verification;
• situations requiring additional editorial approval;
• rules for confidential information, sources, and personal data;
• expectations concerning accuracy, copyright, fairness, and accountability;
• rules governing external AI tools;
• disclosure and labelling of AI-assisted and AI-generated content.
They will also consider how disclosure should differ depending on the role AI played in producing the content, rather than relying on a single label for every use.
4. Create the governance processes
Participants will move beyond the policy document and define how the rules will work in practice.
They will consider:
• who approves AI tools and new use cases;
• who reviews high-risk or sensitive uses;
• how decisions and approved tools should be documented;
• how staff should raise concerns or report problems;
• how the newsroom should respond to AI-related errors or incidents;
• what training staff need;
• how policies, tools, and workflows should be reviewed.
This stage will help them connect editorial responsibility with clear roles and decision-making processes.
5. Test and improve the framework
Participants will apply their draft rules to realistic newsroom scenarios. This will help them identify what documents provides enough guidance when journalists face time pressure, uncertainty, sensitive information, synthetic content, or a new AI tool.
The group will also examine common weaknesses found in AI policies and internal guidelines, including:
• broad commitments without practical instructions;
• undefined terms such as “appropriate” or “responsible” use;
• unclear responsibility and ownership;
• inconsistent disclosure and labelling;
• missing safeguards for sensitive or high-risk reporting;
• insufficient rules for external tools and confidential data;
• no procedure for errors, complaints, or incidents;
• policies that are published but not supported by training or regular review.
