Is AI a threat to Citizen Science?

Citizen science is booming, with projects like the Australian Museum’s FrogID, DigiVol and Australasian Fishes engaging more than 30,000 citizen scientists. This boom, in large part, is due to the engagement participants feel by being directly involved in science and through feeling they are making a difference. Importantly, citizen science has the potential to invoke passion. This passion is the magical ingredient that makes citizen science such a powerful experience for both citizen scientists and for those running the projects.

So what does the future hold for these projects? As technology becomes more sophisticated, high-functioning and complex – a movement often referred to as Artificial Intelligence (AI) – how will citizen science be affected? A recent publication delves into this question; particularly how the intersection of these areas can provide opportunities and risks for citizen science. The publication outlines how AI is already influencing and shaping this field through a range of technologies that can assist or replace humans in completing tasks, possess the ability to influence human behaviour and can help improve insights into data.

These technologies have been incorporated into existing popular citizen science projects such as: iNaturalist, where AI assists in identifying animals and plants from photos contributed by citizen scientists; eBird where AI is used to verify the accuracy and consistency of contributor’s submissions; and CitClops which monitors and provides early warning of algal blooms. AI is also used widely through social media platforms for extending impact of citizen science projects.

There are certainly benefits in the use of AI in citizen science. For example, do you get frustrated or bored by repetitive, routine or mundane tasks causing you to lose interest or tune out from a citizen science project? Do you lack enthusiasm when it comes to certain projects or citizen science as a whole? If so, AI could help. AI may re-engage you in future projects, taking care of the more routine tasks, and allowing citizen scientists to focus more on their interests, or their passions. No longer will you need to tag image after image without any animals in them, as AI could take care of that, leaving you time to focus on more challenging identifications or in observing animal behaviour.

The authors discuss such opportunities as well as drawing attention to potential risks posed to citizen science. Such risks generally revolve around moral behaviour and ethical grounds, whether that be in relation to the humans implementing AI (leading to inappropriate use of AI) or to the mechanisms of the AI itself (resulting in unpredictable and possibly undesirable outcomes). There is also an issue of over-reliance on AI, which can lead to increased error – for example, how often do we rely on something as simple as spell-check and find our actual spelling skills in decline?

Many of us are wary or even downright terrified of AI, as popular media has cast it in a sinister light where machines take over the world. This paper does not verge into that territory. Instead, it offers a balanced view of what AI has to offer citizen science, with examples of real world application whilst acknowledging the risks that we need to consider in order to ensure that the magic ingredients are still in the recipe.

Paul Flemons

Manager Digital Collections and Citizen Science at the Australian Museum