sgibson91
commented
3 years ago Adding notes from meeting with @Ismael-KG
General Tips
- Be explicit when referring to anything: methods, members of the community, etc.
- The more you share/clarify, the easier it is for the committee to review. Show that thought has been put in and justify your decision-making. They don't need the elevator pitch.
- Attach existing documents: proposal, interview script, participant information sheet, consent form, etc.
- Clarify that this research could result in a useful resource for the governance team in the future
- Not an obstruction, but an enhancement
- Project is focused on research and not engaging with processes and policies already in places or currently in development
- Need to acknowledge from the start that not all projects can be open
- Need to protect privacy
- Parts that are made open should be done in a way that facilitates and encourages re-use and remixing
1. Project Goal and Purpose
The Turing's mission statement is to innovate and accelerate data science and AI in the UK. One pathway to achieving this goal includes open working, producing research that can be built upon by anyone, and collaborating with established communities in the open science ecosystem. By interviewing experienced researchers, we will assess the current attitudes towards open working practices within Turing staff, and thereby whether open working is a viable pathway towards innovating the AI landscape. The results of this user research would feed into a wider policy design phase that would encourage and empower Turing researchers to adopt open working practices. These practices will help the Turing advance it's institutional goals:
- Advance world-class research and apply it to real-world problems: By implementing open working practices and producing research outputs that are designed to be built upon, we ensure the work is reusable and applicable across the wider domain
- Train the leaders of the future: So much of the tooling around data science and AI relies on open source software and the communities that contribute to and maintain them. Adopting these practices at the Turing will train our researchers in the skills required for working with such projects.
- Lead the public conversation: In the age of informed consent, the uptake of new methodologies and technologies in heavily reliant on positive public opinion. When the Institute is seen to publish openly, invite community review, and respond well to recommendations, it will garner trust in the Turing's research and facilitate collaborations with other open working communities.
This is ultimately an internal, business-focused research project, but it would be nice to ultimately publish something public to demonstrate what we learned, what we changed, and what effect those changes had in case another institution wishes to replicate such a process.
2. Data and Research Methods Description
-
Type of data: Textual, interview transcripts
-
Amount of data in the sample: Less than 20 interviewees
-
Sources of data: Researchers in the Tools, Practices and Systems research programme, Research Engineering team. Particularly these as they have vested interest in open working embedded into them.
-
Gaining insight from experienced researchers about practices
-
Attach interview script #5
-
Who will be interviewed? How will they be chosen? How will they be interviewed?
-
What preparation has been/will take place to analyse the data
3. Consent
- Develop and attach consent form #8 and Participant Information Sheet #7
- Transcripts will have names/programmes/teams removed and quotes will be modified to maintain anonymity
4. Privacy and Security
- Data will be anonymised with references to people/programmes/teams removed and quotations modified
- Since the data is for internal use only and will be of the form of text documents/transcripts, they will be stored on the Turing's Sharepoint (vetted by Turing IT for security and has access management)
- Raw data such as zoom recordings and automatic transcripts will be deleted after being combined and anonymised
5. Other Harms
- Potentially revealing unhappy working relationships: Contents of interview transcripts will be modified and redacted to anonymise from identification. This process will be outlined in the Participant Information sheet #7 to reassure interviewees and they will be made aware of this before consenting to participation.
- Policy will be to default to open for everything: Clarify from beginning that not all projects can meet this goal for good reasons. Intend to provide a resource that could be useful for future policy and governance questions. There's still an open question around the compatibility of Intellectual Property and Open Licensing.
Summary
Complete the Turing Ethics and Advisory Board's survey to gain go ahead to share data