Beyond Traditional Ethics when Developing Assistive Technology for and with Deaf People in Developing Regions - Ethical Engineering for International Development and Environmental Sustainability - page 308

we write for 'Deaf people' rather than for 'the Deaf'. We always explain what we
mean by 'Deaf' vs. 'deaf'. We try to avoid words like 'target community' and
'human access points' (originally defi ned by Chetty et al. 2004 ) because of their
dehumanising connotations. The Deaf community is not a target; it is a stakeholder
in the action research. Likewise, calling a person an access point, like a piece of
technology, is also not copacetic.

5 How Ethical Issues Affect Technical Design

Most if not all of these issues are dynamic and ever changing, which is part of action
research. This means that roles and projects need to be continually negotiated and
necessitates continual communication between stakeholders. Many of these issues
lie at the fringes or beyond the reach of conventional and traditional notions of
research ethics, as discussed in the opening sections of the chapter. An example is
changing research priorities midstream as dictated by the Deaf community, from
text- to video-based. Does that mean that the ethics approval has to be changed and/
or resubmitted because the media for the consent form and/or data collection has
changed? Another example is challenging the elders in the Deaf community regard-
ing Facebook. Does that warrant explicit acknowledgement in a proposal sent to an
IRB, because it affects primary and secondary data collection by virtue of infl uenc-
ing people's behaviour. These examples highlight why some traditional IRB-driven
ethics considerations may even be irrelevant.
As it turns out, such issues also have ramifi cations for technical design and
evaluation of ICT artefacts, especially the iterative engineering of prototypes. For
example, we moved from text-based to signed language-based prototypes because
of our evolving and deepening understanding of what Deaf people want. Most of the
text prototypes, culminating with SIMBA (Softbridge Instant Messaging Bridging
Architecture), went unused for a number of social rather than technical reasons.
Despite this, the involvement of Deaf people led to a number of innovations. For
example, when we built SIMBA, we used an 'is typing' presence indicator for the
Deaf person, on an instant messaging interface, to represent when a hearing person
was speaking. Because of the lag converting speech to text, and vice versa, we
realised we also needed a similar 'audio is typing' interface for the hearing user and
implemented that with a musical passage.
When we started with video prototypes, our main concern was with video codec
manipulation and sign language intelligibility. However, a simple yet striking piece
of feedback from Deaf participants was that they preferred to have side-by-side
video (see Fig. 10.3 ) instead of picture-in-a-picture, because they wanted to see
themselves more clearly when signing, i.e. the sender's image is too small with
picture-in-a-picture like in a Skype call. We came to this innovation because of
answers to open-ended questions to Deaf participants with the use of an SASL inter-
preter (as noted in Ma 2009 ). Even though we implemented this 50-50 screen split,
i.e. divide a landscape screen into two equally large areas, the idea came from Deaf