By neha.pujari , 19 January 2026

External facing

Allowed
Using AI tools to design the look and feel of publications, including cover designs and infographics.
Category
Publications
Guidance Reference
Ensure appropriate content and media attribution, Ensure responsible use of AI-generated voice, images and videos. Label content created by AI.
Not Allowed
Using AI to copy designs or use prompts such as "in the style of (artist name)".
Use Case Number
UC023
Explanation
AI can enhance the visual appeal of publications, but copying designs can infringe on copyright and ethical standards.
By neha.pujari , 19 January 2026

External facing

Allowed
Allowing the external host to use their own AI solution or a publicly available AI solution to transcribe and summarize meeting notes, recognizing that the information shared in the meeting must be treated as public information.
Category
Meetings
Guidance Reference
Ensuring transparency in AI use, Safeguard WHO data
Not Allowed
Allowing the external host to use their own AI solution or a publicly available AI solution to transcribe and summarize meeting notes, as the meeting will be discussing sensitive and confidential matters.
Use Case Number
UC022
Explanation
AI can help capture and summarize key points from meetings, but recording and using AI tools without consent or sharing sensitive information without safeguards violates privacy and ethical standards.
By neha.pujari , 19 January 2026

External facing

Allowed
A WHO hosted meeting uses approved AI tool to transcribe and summarize meeting notes e.g. Copilot M365.
Category
Meetings
Guidance Reference
Ensuring transparency in AI use, Safeguard WHO data
Not Allowed
A WHO hosted meeting uses an unapproved AI tools to transcribe meetings e.g. Otter.ai; Fireflies.ai
Use Case Number
UC021
Explanation
AI transcription saves time and captures key points, but unapproved tools may expose sensitive information.
By neha.pujari , 19 January 2026

External facing

Allowed
Employing AI for cybersecurity measures, such as detecting and preventing cyber threats.
Category
IT applications
Guidance Reference
Be alert to cybersecurity threats, Safeguard WHO data
Not Allowed
Using AI to access or manipulate external data without proper authorization.
Use Case Number
UC020
Explanation
AI enhances security, but unauthorized access can lead to data breaches.
By neha.pujari , 19 January 2026

External facing

Allowed
Using AI tools, like Midjourney, DALL·E 3, and Adobe Firefly, to create non-realistic images or design elements for public reports, presentations or external communications content.
Category
Image creation
Guidance Reference
Ensure responsible use of AI-generated voice, images and videos, Ensure appropriate content and media attribution. Label AI content
Not Allowed
AI should not be used internally OR externally to create or alter realistic images or videos of people, places or situations.
Use Case Number
UC019
Explanation
AI-generated images can enhance presentations, but misleading images can breach trust and cause confusion and ethical issues. In addition, lack of permissions or attributions can violate copyright laws.
By neha.pujari , 19 January 2026

External facing

Allowed
Employing AI for data analysis to identify trends and insights from external data.
Category
Data analysis
Guidance Reference
Cross-check the output, Follow the WHO Code of Ethics and Professional Conduct
Not Allowed
Using AI to analyze personal data without proper consent.
Use Case Number
UC018
Explanation
AI can improve decision-making and efficiency, but analyzing personal data without consent violates privacy.
By neha.pujari , 19 January 2026

Internal facing

Allowed
AI-powered interpretation tools may be used under an Enterprise Agreement for internal meetings with the WHO workforce only. Hosts can add a disclaimer by announcing it verbally, posting in the chat, or including it in the meeting invite or agenda.
Category
Interpretation tools
Guidance Reference
Practice ethical AI use, Safeguard WHO data
Not Allowed
Using AI-powered interpretation tools not under an Enterprise Agreement for internal purposes, in meetings with WHO workforce only and not adding a disclaimer.
Use Case Number
UC017
Explanation
AI interpretation enhances communication and reach but, due to its simultaneous nature without human review, may pose reputational risks, confidentiality and legal concerns, and potential miscommunication. It should be used cautiously.
By neha.pujari , 19 January 2026

Internal facing

Allowed
Utilizing AI to generate scripts, or support video production, including subtitles, transcripts, storyboards, sound effects, music, voice-over, graphic effects, animations or avatars.
Category
Video creation
Guidance Reference
Check carefully for accuracy and potential bias. Ensure responsible use of AI-generated voice, images and videos, Practice ethical AI use and label content created by AI.
Not Allowed
AI should not be used internally OR externally to create or alter realistic images or videos of people, places or situations.
Use Case Number
UC016
Explanation
AI can make information more engaging, but deepfakes can cause reputational harm and ethical concerns.
By neha.pujari , 19 January 2026

Internal facing

Allowed
Using AI-powered translation tools under an Enterprise Agreement for internal purposes, such as intranet websites, internal training materials, and internal emails. A disclaimer should be included whenever any information is translated solely using AI.
Category
Translation
Guidance Reference
Cross-check the output, Practice ethical AI use, Safeguard WHO data
Not Allowed
Using AI-powered translation tools not under an Enterprise Agreement for internal purposes without review and approval by WHO Language Translation team, such as Governing Bodies Documents, official WHO communications, and Publications.
Use Case Number
UC015
Explanation
AI translations facilitate communication, but it is important to minimizing risks associated with miscommunication, legal issues, and reputational damage. Refer to the WHO Translation Policy in the eManual (I.3.3 and VIII.2.12).
By neha.pujari , 19 January 2026

Internal facing

Allowed
Asking AI-powered chatbots to discover features and answers on how to do certain tasks on application
Category
Support
Guidance Reference
Safeguard WHO data
Not Allowed
Blindly trying the suggestions or answers on sensitive or business critical data.
Use Case Number
UC014
Explanation
Data can be lost on untested or unknown features, especially if there is no rollback or undo mechanisms.