Four ethical principles guide all of CNTI’s work:
- Accountability: As researchers, we are accountable for our work. All authors whose names appear on reports take full responsibility for the work.
- Transparency: As researchers, we explain our work so that our colleagues and readers can understand how we arrived at our conclusions. That also makes it possible for others to replicate our work. All CNTI research contains detailed information on our methods under headers like “how we did this” or “about this study.” We always include information about all resources used for research, including tools, software, vendors and external sources.
- Beneficence: As researchers, we are responsible for minimizing risks and maximizing benefits to participants. For the types of research we conduct, data breaches are the most likely risk. Participants’ information is anonymized before it is shared outside the research team, including with AI tools. The process varies with the methods but can include the use of pseudonyms, data aggregation, and other procedures that protect the privacy of research participants.
- Independence: As researchers, we are independent of outside interests. We declare our funding sources and ensure that there are no conflicts of interest in our work. Our professional staff make all consequential decisions.
The same principles guide our use of AI tools.
CNTI uses AI in the following ways for communications: to generate images for our website and print materials, which are denoted with the “noise” filter and associated labeling; to generate summaries, which are reviewed and edited by humans; to translate the text on our website from English into other languages, which is done using Google Translate from the button on the top right of CNTI’s homepage; and to generate AI “podcast” versions of our newsletter, which are labeled as such.
CNTI uses AI in the following ways for research: to translate materials for internal review, which are reviewed by human translators before they are included in published work; to transcribe audio recordings, which are reviewed by the research team; to assist with grammar and readability when writing; and to assist with brainstorming, which is always subject to human review and verification.
Share
