Muck Rack's AI standards

Introduction

AI (including generative AI, machine learning and data science) is changing the way we work. At Muck Rack, we believe professionals in PR and journalism must embrace AI thoughtfully and responsibility. Generative AI is a productivity tool that should be used to help professionals save time on repetitive and tedious tasks so they can focus more on strategic and creative work. This standards guide serves as a living document of best practices that will be updated as the technology evolves.

Confidentiality

• Information inputted into generative AI prompts should be mindful of confidentiality and trademark laws and free of specific names, proprietary or confidential information and copyrighted works.

Responsibility and accuracy

• Content created using generative AI requires human oversight, and must be scrutinized and fact-checked the same way human-created content is.
• PR professionals and journalists are responsible for the content they create and distribute, whether human- or AI-created, and must check it for accuracy, fairness, confidentiality and originality.
• Images created using generative AI should be labeled as such.

Management and training

• It’s the responsibility of an organization, whether media, agency or brand, to create, apply, enforce and update their own ethical guidelines for when and how AI should be used, how to disclose its use, and how content created with generative AI should be fact-checked and edited.
• Management teams must determine the acceptable applications of AI and distribute guidance and regular training to their staff, whether for a newsroom, content marketing team or other setting. For example, a social media post, story outline, or SEO-optimized headline may be acceptable, while using AI to draft an article or generate an image may not be.
• Management teams should establish a review process for investigating questionable AI practices.

PR Agencies

• In general, agencies using AI responsibly and following the above best practices when integrating AI into their workflow do not need to disclose its use to clients, unless there are concerns around confidential information being shared or are contractually obligated to do so.
• It’s best practice to be forthcoming and transparent with clients, and agencies should be aware of clients’ privacy and liability concerns before inputting confidential information into AI tools.