Ethical Issues

It is important that students consider the potential ethical issues associated with their projects and this has been built in to the project process. Students should discuss potential ethical issues with their supervisor to determine whether they need to be addressed and, if so, how they should be dealt with. As a general rule of thumb, ethical issues should be considered whenever human participants are involved in the project. More information and guidance can be found on the University of Leeds research ethics web page.

Under the University's guidelines, each project must be ethically approved. This means that either it must be agreed that there are no ethical issues to address or that it is covered by block ethical approval granted by the Faculty Research Ethics Committee. The School of Computing has been granted block ethical approval for 5 types of project activity, details of which are given below. If you are concerned that your project will be outside these categories then you should contact the Projects Coordinator.

In order to ensure that ethical issues are addressed appropriately, the following procedures will be carried out as part of the project process.

1. Aim and Minimum Requirements: feedback on these from the projects coordinator will include highlighting potential ethical issues. The supervisor will advise the student on how these issues may be addressed appropriately.
2. Mid-Project Report: this will include a compulsory appendix ``How ethical issues are addressed''. Feedback from the project assessor will include whether ethical issues, when they arise, are being addressed appropriately.
3. Progress Meeting: if any ethical issues have arisen during the project the student, supervisor and assessor will agree at this stage on how this should be demonstrated in the final report.
4. Final Project Report: each student will include a compulsory appendix in their final report, entitled ``How ethical issues are addressed''. This will include, when appropriate: copies of emails from data owners; a description of how consent was obtained from participants; consents from web-site owners; measures taken to preserve anonymity in images; ethical assessment for NHS data.

An introductory ``Procedures and Timetable'' meeting for all undergraduate project students will include discussion of ethical issues. Additional lectures will be provided to illustrate potential ethical issues relating to students in the School of Computing.

Block Ethical Approval

Block ethical approval has been obtained from the Faculty Research Ethics Committee for project activities which can be classified as one of the following five categories.

• Images of People:
  
  Projects in the field of Computer Vision often involve the analysis of still and moving images of people to detect features or recognise behavioural patterns. For example:
  • Tracking individuals in crowded scenes: this has applications in security (analysis of CCTV footage) and sports analysis.
  • Facial feature recognition: this includes recognition of facial expressions.
  Still and moving images of people are used here to evaluate and improve the algorithms and software being studied by the student. The images used are typically taken from purpose-built datasets used within the research community or selected from publicly available images on the WWW. Projects involving tracking individuals might also use CCTV footage or new footage filmed under controlled conditions by the students.
  
  In cases where new footage is being filmed, participants who are recruited to be the focus of attention in the new filming will be provided with an information sheet, describing the purpose of the project, the scenario that is to be filmed and the use to which the images will be put (see the accompanying template, the details will vary from project to project). They will also be asked to provide written consent (see the accompanying template for a written consent form).
  
  The main ethical issues are associated with anonymity and, for existing images, the legitimacy of the source. In the use of still and moving images of people, the student will ensure that all images are taken from legitimate sources (guided by the supervisor), that they are permitted to use the images for the project, and that copyright laws are adhered to. In addition, if CCTV footage is used or new footage is filmed for the project, the student will carry out an impact assessment to establish whether the images can be gathered in such a way that individuals cannot be identified, e.g. by not pointing the camera at faces, or by pixellating faces. If footage is being recorded in a public area, passers-by will be made aware of the filming by putting signs up and an alternative route which avoids the filming will be pointed out. Unless written consent has been obtained from an individual to use his/her image without anonymisation, any images collected in this way which appear in the final report will be anonymised, e.g. by pixellating faces.
  
  
• Medical Images:
  
  Projects in the field of Computer Vision and Visualization often involve the application of algorithms and software to medical images. For example:
  • Automating the detection of cancerous tissue: cancer cells look different to healthy cells so computer vision techniques can be used to identify cancerous tissue. Such techniques can also be used to classify tissue samples as healthy or not.
  • Development of an interface to help in the visualization of histopathological data (images of tissue samples).
  The images are used here to evaluate and improve the algorithms and software being studied by the student. The aim is to develop tools which can help to analyse images of tissue samples. The images used must be fully anonymised.
  
  The main ethical issues are associated with anonymity and the legitimacy of the source. In the use of medical images, the student will ensure that all images are taken from legitimate sources and that they are permitted to use the images for the project. All images used will be anonymised before the student has access to them. The supervisor will be able to verify the legitimacy of the source.
  
  
• Use of Religious Texts:
  
  Projects in the field of Natural Language Processing typically involve the automated analysis of text, either for translation purposes or to find patterns which will relate documents or parts of documents in a large corpus. These documents are sometimes religious texts. For example:
  • Machine translation of the Quran in to other languages.
  • Categorisation of verses of the Quran.
  
  The main ethical issues are associated with ensuring sensitive use of religious texts. Copyright laws must also be complied with. The student, guided by the supervisor, will ensure that religious texts are used sensitively, to avoid causing offence, and that copyright laws are complied with.
  
  
• User Testing:
  
  Projects in Computing typically involve the design, development and/or evaluation of software tools and/or human-computer interfaces. In order to obtain objective feedback on the software the student may enlist a selected group of potential users to test the software. For example:
  • Testing the usability of a graphical user interface.
  • Evaluating whether a visualisation technique helps people to understand and interpret data more easily.
  The user will typically be asked to perform specific tasks using the software (e.g. extracting specified information from a database). These tasks may be observed by the student to determine the ease with which they are performed. The user will also be asked to respond to questions relating to ease-of-use, functionality, potential improvements. These will most commonly be presented in the form of a questionnaire, but may involve semi-structured interviews.
  
  If testing is to take place at the workplace of an external company then the student will make sure that they have been briefed on the local health and safety protocols.
  
  Students will identify appropriate ``users'' under the guidance of their supervisors. The choice of users will depend on what the software being tested is designed to do and may require a mix of experts and non-experts (with basic IT skills). E-mail may be used to recruit potential participants. When it is appropriate for fellow students to act as testers then they may be recruited face-to-face. For web-based software, which may be evaluated remotely, potential users may be identified and recruited through appropriate on-line forums.
  
  The student will provide each participant with an information sheet which describes the purpose of the project, the data that will be collected and the use to which the collected data will be put. The participant will be assured that they can stop and withdraw from the user-testing at any time. Please see the accompanying template information sheet, the details of which will vary from project to project. When a questionnaire is used to collect data, the participant's consent is implicit in the completion of the questionnaire. When appropriate, the student will also provide a written consent form (see accompanying template).
  
  The main ethical issues are associated with avoiding coercion of participants and ensuring anonymity. The users selected by the students will very often be fellow students. In order to avoid coercion the student will ensure that the testers/participants are aware of their rights, and that they know that participation is voluntary. No sensitive information will be recorded and any personal information collected will be stored securely.
  
  
• Use of Data from the Internet:
  
  Some Computing projects involve the analysis of data collected from the WWW. For example:
  • Publicly available statistical datasets, such as the UK government national statistics, may be analysed using visualisation or pattern recognition techniques to identify trends and correlations.
  • Publicly accessible websites may be used to provide documents which can be used to evaluate data mining techniques, e.g. discussion forums for online technical support.
  • Data about web site usage, such as how often they are accessed, may be analysed to extract trends.
  This data is typically used to test/evaluate computational algorithms in information visualisation, natural language processing, informatics.
  
  The main ethical issues are associated with consent and anonymity of data. Where necessary, the student will obtain consent from the website owner to collect and use the data. All data collected which would allow an individual to be identified will be anonymised.