What are you looking for?

RDA Guidance on AI Tools Usage

Version: July 2025

The updated RDA guidance extends beyond Artificial Intelligence (AI) meeting assistants to cover different types of AI tools used in RDA meetings, events, and content creation, providing comprehensive guidelines for responsible and ethical AI use across Research Data Alliance activities.

1. Introduction

The Research Data Alliance (RDA) AI Usage Guidance has been developed in response to the growing adoption of AI tools by researchers and academics for various activities, including literature reviews, writing assistance, software development among others. Recognising that AI tools are increasingly integrated with web conferencing platforms to enhance productivity during work-related tasks and meetings, RDA acknowledges both the vast potential these tools offer for Open Science and the need for community guidance to ensure their responsible implementation within RDA activities. The guidance takes a pragmatic approach recognising how AI tools can help productivity, and in some cases inclusivity (text-to-speech, live closed captions and translations, note-taking help for neurodivergent members etc.) while remaining mindful of both the advantages and disadvantages these technologies present.

2. Scope

Since AI encompasses many different technologies that are fundamentally based on data, which directly aligns with RDA’s core focus on data-driven innovation, this guidance applies universally to all community members participating in RDA activities, including meetings, events, and content creation, regardless of membership status. The guidance specifically addresses AI tools that create new content, including AI meeting assistants that record audio, transcribe notes, capture slides and screenshots, and generate summaries. 

With the primary goal of supporting responsible use while minimising risks of misuse that could harm individuals or violate RDA’s mission and guiding principles, the guidance emphasises core Open Science values of transparency and reproducibility by providing best practices for members to follow. Given the rapidly evolving nature of AI technology, this guidance is designed as a living document that will be regularly reviewed and updated to remain current and effective in addressing emerging challenges and opportunities.

3. Context

Text rendering of spoken discussions and presentations provided by AI meeting assistants may help to promote inclusion, accessibility and participation during meetings by assisting those with disabilities, not fluent in the language used, and/or unable to fully engage for the entire meeting duration. Additionally, GenAI tools can provide real-time translation, create accessible summaries, and offer writing assistance that further supports diverse participation needs. However, there are privacy implications since some AI meeting assistants use the data collected to improve their transcription services (e.g. read.ai), and GenAI tools often require extensive data processing by third-party providers with varying privacy standards. As such, while AI tools can assist participants, RDA recognises it may also cause the participants to feel unsafe and unwilling to speak freely, particularly given concerns about how their contributions might be stored, analysed, or used to train AI systems.

4. Guidance

In alignment with the RDA’s Code of Conduct, the following guidelines and best practices aim to provide a safe and inclusive environment for all community members and meeting participants:

4.1. Due Diligence

Understanding tools and their limitations

  • Tool Selection: Choose tools with open and transparent methodologies and documentation, familiarising yourself with privacy policies where available. 
  • Accuracy Awareness: Recognise that AI tools may generate biased, inaccurate, or inappropriate results and understand potential causes based on data usage policies.
  • Quality Assurance: Ensure accuracy before publishing any AI-generated information. 
  • Content Responsibility: Take responsibility for accuracy of published information, including AI-generated content. 
  • Citation Requirements: Follow appropriate citation guidelines for AI tool usage (see this example).

4.2. Consent and Meeting Procedures

Obtaining participant consent and managing meetings

  • Pre-meeting Notification: Participants must notify the meeting organiser/chair about their intended AI tool usage and explicitly inform all participants to request their consent for using the tool before meetings.
  • Meeting Declaration: Organisers/chairs must announce any AI tool use at the beginning of the meeting and proceed with the tool usage only if all participants give their approval.
  • Bot Restrictions: Organisers/chairs must prohibit AI bots from attending meetings as participant substitutes. This violates RDA ‘consensus’ and ‘community-driven’ principles, and breaches the RDA Code of Conduct (inappropriate content, “Opinions of fictitious parties”).
  • Facilitator Authority: Enable meeting facilitators to identify, declare, and remove undisclosed AI bots as necessary. 
4.3 Data Privacy

Protecting Participant Information

  • Clear Privacy Statements: Use only tools with transparent data privacy and usage policies detailing collection practices.
  • Policy Examples: Reference privacy policies (e.g., Otter.ai, Sembly.ai, Fireflies) that demonstrate security and compliance best practices. See this example guide on privacy settings and ‘Erasing AI memory’ for the most common Generative AI tools like ChatGPT, Co-pilot and Claude.
  • Personal Information: Sharing personal, sensitive or confidential information with AI tools is not allowed. Make sure to remove personal data.

5. Shaping Responsible AI with Data: RDA’s Perspective 

RDA recognises the immense impact and influence that AI has across research domains, understanding that like any powerful technology, it carries significant ethical implications where the benefits may not always outweigh the challenges. 

As a data-centric organisation, RDA aims to emphasise the critical importance of high-quality data as the foundation for these technologies, recognising that the RDA community is uniquely positioned to play a pivotal role in shaping how AI develops in terms of applications, outreach, and impact. At the current state-of-the-art, data represents one of the most valuable resources and barriers to access alongside infrastructure, making the community’s collective expertise invaluable. 

Regardless of individual AI expertise levels, RDA community members possess deep knowledge about all aspects of data, presenting an unprecedented opportunity to actively participate in shaping this technology for good rather than remaining passive observers, ensuring that the data foundations underlying AI systems support ethical, inclusive, and beneficial outcomes for the research community and society at large.

We encourage all RDA members to actively contribute to this guidance by sharing their experiences, practices, and reflections on AI tool usage within RDA activities or the broader research context.

Please contact [email protected] with your feedback and suggestions. 

Disclaimer: This version of the RDA Code Guidance on AI Tools Usage was updated by the RDA Secretariat on 14 July 2025. AI (Claude.ai Sonnet 4 Pro) was used to improve the language and tone of the text and has been verified and edited.