In our latest publication through the HARMONIZE consortium, researchers have collated evidence and synthesised knowledge from 80 full-text sources. The research aimed to get an overview of what ethical considerations have been mentioned in the context of drones used for research in the fields of environment and health. In our methodological approach, we considered ethical behaviour through four distinct lenses (see Figure 1):

1. Research Conduct: How are drone studies conducted and reported on in environment and health research?
2. Communities: How do local communities and the public perceive drone research?
3. Research Reflection: What concerns and mitigations do researchers reflect on?
4. Guidance: What ethical guidance has been made available to researchers?

Figure 1. The four central research questions of the scoping review.

Our methodology has been previously described in our research protocol. We screened both academic databases and grey literature, going through 6426 abstracts, 290 full-texts, and finally selecting 80 full-text sources for inclusion. These 80 sources were then categorised into one or more of the four groups, which is why the mindful reader will notice the numbers in Figure 1 do not add up to 80!

Information from literature was extracted into tables that provide a useful resource for researchers working with drones in the field, especially when designing a new study. By checking the Research Conduct table, they can, for example, learn what common drone applications are, how these studies are designed, and most importantly, how they have organised community engagement and flight protocols in that study's setting. Furthermore, in the Communities table, the actual perceptions measured by qualitative studies or surveys conducted within local communities and the wider public provide insight into what actually matters most to those living in areas where research is being done. The results for the Research Reflection category were analysed through thematic analysis, where researchers systematically identify, analyse, and report recurring themes that occur throughout the full-texts. To supplement the literature, an expert advisory workshop was held in Barcelona to ensure no themes were missed. Finally, although we did not find many publicly available ethical guidelines (only 4) for drone researchers, the scope of these guidelines was summarised in the last table of our publication, making it easier for researchers to find these and to spot any gaps in the currently available guidelines.

Figure 2. Graphical Abstract.

One takeaway for researchers designing their next drone study is that the easiest way to improve community engagement is to include qualitative study components from the start. There are many ways to do this, including, but not limited to: action research, citizen science, focus groups, participatory mapping, transect walks, interviews, sketch mapping, or otherwise interactive workshops where, for example, researchers construct drones with members of the community or hold practice flights.

The thematic analysis showed that many drone researchers have already actively thought about how to make their research live up to higher ethical standards. We grouped the concerns and mitigations they voiced into seven groups: social impacts, privacy, regulatory, autonomy, beneficence, safety, and technology. Privacy is the biggest concern, about which researchers express nuances between data privacy (who owns and can share the data), the feeling of being observed, what is a (legally) reasonable expectation of privacy in different contexts (own home vs public land), how should incidental findings (unintentional recordings) be handled?, and how should we deal with the increased risk of individuals being personally identifiable through the recording of their image or belongings due to the high-resolution of many drone cameras?

There were very few studies that directly asked community members about their experiences with drones being flown in the area where they reside. The answers they provide on what they consider to be sensitive or private information are highly specific to their socio-cultural contexts, highlighting the need to speak to members of the community directly and not rely on assumptions based on previous experience in other research areas. The majority of studies were online surveys that interrogated the public on their attitudes and knowledge of drones. Overall attitudes towards drone use for scientific research were positive and better for other applications that were not viewed as benefiting the public good (e.g. commercial use).

Through our search, we were only able to identify four ethical guidelines that specifically covered the use of drones in environment and/or health research. Most of the focus was placed on technical compliance and safety issues. While these are, of course, important, as they will have large consequences if not adhered to, more attention could be directed towards community engagement strategies.

The results from this scoping review can be used to update and complement existing guidelines. We also observed a gap in the literature when it comes to documenting the community experiences of drones being flown within an area. Additionally, some of the tougher dilemmas remain unaddressed, such as how to use drones ethically in areas with ongoing or a history of conflict (e.g. war, or illegal activities), or what to do when you assume an area is uninhabited, but yet you find you have recorded individuals upon checking the recording at a later date? Some technological improvements may be able to mitigate these in the future through privacy-by-design features or the incorporation of AI algorithms that could shut down cameras automatically when unintentionally recording personally identifiable information. In any case, responsible research starts with early and meaningful community engagement during the research conceptualisation phase. This is almost never a linear and predictable process, it requires constant communication and updating of the researchers' understanding of the local context they are working in.