At a session devoted to ‘science and society’ on 5 March, UNESCO’s Natural Sciences Sector joined forces with the Social and Human Sciences Sector to examine the implications for society of modern applications (apps) which incorporate artificial intelligence to improve environmental monitoring. The 90-minute session asked whether apps designed for environmental monitoring were fostering citizen scientists or whether they should be considered just another consumer product?
The session was organized as part of Mobile Learning Week at UNESCO headquarters in Paris, which is focusing, this year, on the theme of artificial intelligence for sustainable development.
In front of an audience comprised mainly of young adults, Dr Will Logan presented free rainfall and climate apps that can be used by individuals, schools, local governments and national hydrological services alike to estimate rainfall. These apps provide estimates using remotely sensed data transformed by artificial intelligence (artificial neural networks). The apps are user-friendly and, in the case of the iRain app, take a ‘citizen science approach’ by requesting input from the stakeholder or citizen to improve near real-time rainfall estimates. Thanks to the presence of geostationary satellites, the apps can provide estimates with a time lag of just 45–60 minutes.
Dr Logan demonstrated how the apps could be used to anticipate water-related hazards such as flooding, droughts and storms. He mentioned that there were thousands of users around the world, including meteorological services, wildlife parks, ecolodges and large farms. He had found that people were also tailoring the apps to their own needs and finding uses for them that the developer had never imagined. An app could be used to track rainfall over the past 35 years, for instance, to see whether climate change had affected precipitation over a given area.
The applications presented by Dr Logan have been developed since 2005 by the University of California, Irvine (USA), together with UNESCO’s International Hydrological Programme, within the G-WADI programme, and by the institution headed by Dr Logan, the International Center for Integrated Water Resources Management (ICIWaRM) in the USA, which functions under the auspices of UNESCO (category II institute).
Dr Logan was followed by Dr Régine Vigne-Lebbe, Professor of Biodiversity informatics at the Sorbonne University in Paris. She presented various apps developed by her university together with the Museum national d’Histoire naturelle (National Museum of Natural History) to broaden the public’s environmental awareness. These apps help people to identify a particular plant or animal species.
The first category of these computer-aided identification systems are connected objects. ‘With these, a person observes a specimen then transcribes what they see using terminology displayed by the app’, she said. The second category uses artificial intelligence (AI) to provide automatic identification of specimens, such as plants. ‘This type of AI system is already available for some applications’, she added.
Dr Vignes Lebbe cautioned budding citizen scientists to remain vigilant when using the second category of app. Machines excelled at data processing and at compiling statistical records, she said, but deep learning worked like a black box. The machine did not explain how it arrived at a certain conclusion, so the user did not learn how to recognize future specimens.
The first category of app, on the other hand, used verbal communication and did not provide an automatic identification process. Rather, these apps offered advice and an explanation of how they arrived at their findings, enabling the user to learn the right terminology for describing a plant or animal. Armed with this knowledge, the user would be better equipped in future to identify a species and consult literature on biodiversity.
Dr Vigne-Lebbe predicted that future applications would combine human active observation with automatic recognition for optimal results. She described some potential scenarios and insisted on the educational role these apps could play.
The third panelist, Dr Vanessa Nurock, Associate Professor in Political Theory and Ethics at Paris 8 University, examined apps from a philosophical perspective. Apps like those described by Drs Logan and Vigne-Lebbe were great because they were changing people’s lives in many ways – in some cases, they might even save lives, she said – but they also entailed sacrifices.
These sacrifices might be obvious, such as the amount of time a person might spend using the app or their decision to relinquish their right to privacy by giving the app their geographical location. Sacrifices might also be hidden, such as when the apps tamed the user, limiting their options. The waste generated by the energy needed to store and process data and by the mobile phone or computer itself was another hidden sacrifice.
These sacrifices could turn us not only into consumers but also into clients, in the Roman sense of the word. ‘In Ancient Rome’, she explained, ‘a free client was completely dependent upon a patron, a state that was only one step away from slavery’.
Today’s ‘client’ tended to develop a relationship of dependency on the satellites and cloud used by the app by accepting as the ultimate truth the information that he or she retrieved from them. In this way, apps had become the modern equivalent of the Gods or oracles worshipped in the past, Dr Nurock observed.
‘These apps are changing us’, she said. ‘It is important to be aware of the sacrifices we make to these new gods’. However, she added, ‘it is not enough to analyse these sacrifices on the basis of a cost/benefit ratio. We must also analyse them using two other conceptual tools’.
The first of these tools is the ‘apparatgeist’, or spirit in the machine, a term coined by James Katz from the University of Boston. The term could be translated as ‘the spirit in the machine’. It refers to both the design and use of the technology. ‘Apps are not just regular tools’, Dr Nurock explained. ‘They may become a part of ourselves and change our relationship to the world and to others’.
‘Secondly, it is important to bear in mind the distinction between freedom from something and the freedom to act. For instance, designing a system that is not entirely dependent on AI by keeping the human in the loop guarantees freedom from AI. However, this does not mean that citizens are able to understand and control the system, that they have the freedom to stay actively in the loop’.
Dr Nurock concluded that ‘this is why designing an AI system that truly empowers citizens – and allows them to appropriate the app – is so important. As stated in the programme of Mobile learning week, we must enable an AI-literate citizenry when developping apps. I would add that we have a shared responsibility to do so!’