Artificial Intelligence Policy
We use artificial intelligence to help make our work more useful, accessible, and efficient. Our journalism remains human-led, human-edited, and human-accountable. We do not use artificial intelligence as a substitute for reporting, editorial judgment, or accountability.
Our use of AI is guided by our commitment to trust with the communities we serve, and our belief that technology should strengthen, not weaken, the value of our work. We are further guided by our Ethics Policy.
Any use of artificial intelligence must meet the same standards for verification and editorial review as any other part of our work. When we make a mistake, whether in fact or in context, we correct it in accordance with our Corrections Policy.
We are careful about privacy, confidentiality, copyright, and source protection when using artificial intelligence tools. We do not enter confidential or unpublished sensitive information into unapproved AI systems. We are further guided by our Privacy Policy.
What we may use AI for
We may use AI to assist with tasks such as:
- analyzing large datasets or documents;
- assisting with research, fact-checking, and verification;
- brainstorming headlines, questions, or story structure;
- detecting errors in spelling and grammar;
- editing photos, audio, and video;
- generating first drafts of our written work, including sponsored content;
- summarizing audio, visual, and/or written work from other sources;
- transcribing, translating, and making information more accessible;
- product, coding, workflow, and business operations work.
When we use AI in these ways, a human remains responsible for reviewing the work, checking it for accuracy, and deciding whether and how it will be used.
What we do not use AI for
We do not allow AI to replace reporting, editorial judgment, or accountability. We do not use AI to produce original reporting for publication on its own. We also do not use AI to create misleading images, audio, or video that could cause readers to mistake synthetic material for authentic reporting.
When we disclose AI use
We will disclose our use of AI when it plays a significant role in published work, especially when:
- AI was used to analyze material at a scale that would not otherwise have been practical;
- AI was used to generate images, audio, translation, or other content presented directly to the audience;
- a reader could reasonably feel misled if we did not explain how AI was used.
If you would like more information regarding our Artificial Intelligence Policy, please contact hello@taprootpublishing.ca.
Frequently Asked Questions
Why does Taproot use AI at all?
We use AI because it can help us do some parts of our work more efficiently, especially repetitive or time-consuming tasks. That can give our team more time for the work that matters most: reporting, editing, verifying, explaining, connecting dots, and serving our community.
Our goal is not to replace human judgment. It is to use technology carefully in support of Taproot's mission: helping communities understand themselves better.
Does AI write Taproot stories?
Sometimes AI helps produce a rough first draft. But we do not ask a chatbot to "write an article about X" and then publish what comes back.
When AI is involved, it works from material selected and structured by Taproot, such as reporting notes, transcripts, source documents, or other inputs. AI does not assign, report, verify, edit, or publish Taproot stories. People do.
Learn more about how we use AI in our work at our blog.
Do humans verify everything before publication?
Yes. Everything we publish is reviewed by a person. We do not treat AI output as authoritative.
AI tools can be useful, but they can also make mistakes, omit context, invent details, reproduce bias, or present uncertain information too confidently. Our responsibility is to verify information using appropriate sources, editorial judgment, and human review.
Do you use AI to generate images?
No, we do not currently use AI to generate images for our journalism.
There may be appropriate uses for AI-generated illustrations in the future, especially to help explain complex ideas. But we are not using them right now. If we ever do, we will disclose that clearly and take care not to mislead readers.
Do you use reader/member data in AI tools?
We do not use private reader or member information, such as account details, payment information, email addresses, or membership records, to generate journalism.
In some cases, we may use AI tools to help analyze information that readers or members have voluntarily submitted to Taproot, such as anonymous survey responses. For example, during our 2025 municipal election project, we used AI to help identify themes and patterns in anonymous responses to our election question.
When we do this, we use AI to support analysis, not to make final editorial decisions. We disclose to readers when we use AI in this way, and we are careful to protect privacy and confidentiality in accordance with our Privacy Policy.
Do you train AI models on Taproot content?
No, we do not currently train our own generative AI model on our journalism.
We also do not intentionally provide our journalism to third-party generative AI systems for model training. Like many publishers, we continue to monitor how AI companies use publicly available web content.
Will AI replace Taproot journalists?
No. Taproot's value comes from human judgment, local knowledge, relationships, curiosity, care, and accountability. AI can help with some production and research tasks, but it cannot replace the work of understanding a community, asking better questions, building trust, or making editorial decisions.
We see AI as a tool that can support our team, not as a substitute for journalists.
What happens if AI contributes to an error?
If we make an error, we are responsible for it. That is true whether AI was involved or not. If AI contributes to a mistake in something we publish, we will correct the error in accordance with our Corrections Policy. We will also review how the mistake happened and adjust our practices where needed.