Inclusive and accessible science: citizen science in the age of AI
Fotis Mystakopoulos
Aug. 11, 2025, 11 a.m.
Image attribution: Alan Warburton / https://betterimagesofai.org / Image by BBC
We are in an era where Artificial Intelligence (AI) is influencing many scientific fields. Researchers are exploring how AI can be utilised to advance scientific research—while also ensuring it is used responsibly. As part of this ongoing dialogue, the ECSA Citizen Science and Open Science working group organised a webinar to explore various practices demonstrating the potential of citizen science to interact with AI.
What does the audience think?
The webinar began with an exercise to gather participants’ perspectives. In response to the question “How do you use AI in your work?”, the most common use of AI tools was for content generation—text or images. Data analysis was another popular use, and last but not least, participants reported using AI to create or review code.
The discussion continued with a question about where AI adds the most value to citizen science. The most frequent response was data analysis, followed by feedback—though the type of feedback was not specified. Some unique responses, quoted verbatim, included: Improving accessibility through machine translation; I do not see the value yet. I think CS has a lot to add to AI though!; Inclusion and science communication. These responses indicate that there is a perception that AI could facilitate more inclusive and accessible engagement with science.
Finally, participants were asked about their concerns regarding AI in citizen science. Integrating AI into such projects raises critical issues around data quality, potential biases, and the risk of diminishing citizen engagement and skill development. Ethical considerations—including privacy, data security, and responsible AI use—must be addressed to ensure that these powerful tools support rather than hinder the contributions of citizen scientists. While participants expressed more consensus on the value of AI, their concerns reflected a broader and more diverse range of perspectives.
Let’s take a closer look at the presentations from the webinar.
Leveraging the collaborative power of AI and citizen science for sustainable development
Dilek Fraisl’s presentation, “Leveraging the collaborative power of AI and citizen science for sustainable development,” explored the reciprocal relationship between artificial intelligence and citizen science to better monitor and achieve sustainable development. She highlighted how AI is already enhancing citizen science—automating species identification, improving data quality, and boosting user engagement through real-time feedback. Emerging tools like chatbots and conversational platforms offer further potential to support citizen science projects and citizen scientists. Importantly, Fraisl stressed that citizen science can shape AI. By contributing local, context-specific data—especially from underrepresented regions—citizen science can help address gaps and biases in AI models. She pointed to an example from Ghana, where an AI tool developed in Europe was not able to identify water sachets due to the absence of relevant training data from the country. Citizen science can also promote fairness and inclusion by supplying more representative data and involving communities in co-creating ethical AI solutions. Fraisl connected these insights to broader Open Science principles, underlining the need for accessible, transparent, and privacy-respecting data to enable responsible AI. Her talk positioned citizen science as both a user and a shaper of AI—demonstrating its critical role in ensuring AI serves societal needs sustainably and equitably. Fraisl also highlighted the limitations of both AI and citizen science, emphasising the need for careful integration of the two for monitoring and achieving sustainable development.
Co-creating AI for societal challenges – the 'amai!' approach
Karen Verstraelen presented the amai! project in her talk “Co-creating AI for societal challenges – the ‘amai!’ approach,” showcasing a citizen-driven initiative from Flanders that places public participation at the heart of AI development. Rather than following a purely technology-led path, amai! (Flemish for ‘Oh, amazing!’) engages citizens from the earliest stages of idea generation through to the development and implementation of AI applications. With a clear focus on societal challenges—such as mobility, climate, health, and work—the project ensures that AI solutions are grounded in real-world needs and values. amai!’s structured four-phase engagement process includes collecting citizen research questions, co-creating with experts, launching open project calls, and funding AI applications that embed citizen science. Alongside this, the project places strong emphasis on AI literacy, using creative outreach tools like theatre, games, and interactive exhibits to engage broader audiences. Verstraelen highlighted how amai! not only integrates citizen science into AI development but also creates pathways for citizens to influence and shape AI directly. The example of the Waste Watchers project—evolving from a public idea on litter monitoring to a drone-based initiative using AI for image recognition—illustrates this collaborative model in action. As amai!looks ahead, it continues to refine its approach with a view to long-term sustainability, reaffirming the value of citizen co-creation in building ethical and inclusive AI.
Harnessing AI to accelerate citizen science in neuroscience and beyond
Etienne Serbe-Kamp explored the transformative role of artificial intelligence in citizen science through a series of hands-on neuroscience projects, as presented in his talk “Harnessing AI to Accelerate Citizen Science in Neuroscience and Beyond.” He demonstrated how large language models (LLMs) can accelerate research by supporting code writing and data analysis—tools that empowered school students in Chile to investigate electrical signals in plants and authoring a scientific publication (Madariaga et al. 2024). Beyond accelerating workflows, AI was central to creating innovative educational tools, such as the “Computational SpikerBox,” which simulates neuronal activity and was developed using AI-generated code and learning materials. His work also showcased AI’s potential in more complex data analysis, such as building regression models to study public perceptions of climate scenarios. Underpinning all of this was a foundation of accessible, low-cost infrastructure that enabled meaningful participation. While the focus was on how AI enhances citizen science, Serbe-Kamp also pointed to the reciprocal potential—citizen scientists helping to refine AI models through their input—underscoring the collaborative nature of this evolving relationship.
AI in need of citizen science
The webinar stressed the importance of Open Science in the relationship between citizen science and AI, showing how openness and transparency are key to unlocking their full potential. For AI to work effectively with citizen-generated data, that data must be accessible, findable, and reusable—principles at the core of Open Science. Dilek Fraisl noted that while openness is vital, it remains a challenge within citizen science itself. Etienne Serbe-Kamp’s work demonstrated how existing open infrastructures can support AI integration. Open Science also helps guide how data is created and reused, especially in diverse local contexts. Ultimately, the usefulness and integrity of AI depend on strong, collaborative data practices between scientists and citizens alike.
While it may sound bold to claim that AI needs citizen science, the webinar highlighted the essential role citizen science plays in developing more ethical, inclusive, and effective AI. Across presentations, a clear message emerged: citizen science not only benefits from AI but actively shapes it. By contributing local, often underrepresented data and lived experience, citizen scientists help address gaps and biases in AI systems. Projects like amai! show how public engagement can guide AI development, while examples from neuroscience highlight how citizen input can refine AI tools in practice. As the dialogue deepens, co-creating AI with citizens is no longer optional—it’s essential for building more equitable and responsive technologies.
You can watch the full recording of the webinar here: https://www.youtube.com/watch?v=1iYWuyYR4PY
Further Reading:
Duerinckx, A., Veeckman, C., Verstraelen, K., Singh, N., Laer, J. V., Vaes, M., … & Duysburgh, P. (2024). Co-creating artificial intelligence: designing and enhancing democratic ai solutions through citizen science. Citizen Science: Theory and Practice, 9(1). https://doi.org/10.5334/cstp.732
Fraisl, D., See, L., Fritz, S. et al. Leveraging the collaborative power of AI and citizen science for sustainable development. Nat Sustain 8, 125–132 (2025). https://doi.org/10.1038/s41893-024-01489-2
Madariaga, D., Arro, D., Irarrázaval, C., Soto, A., Guerra, F., Romero, A., … Marzullo, T. (2024). A library of electrophysiological responses in plants - a model of transversal education and open science. Plant Signaling & Behavior, 19(1). https://doi.org/10.1080/15592324.2024.2310977
Serbe-Kamp, É., Fraisl, D., & Verstraelen, K. (2025, May 6). Exploring the Role of Artificial Intelligence in Transforming Citizen Science: Privacy, Data Sharing, and Open Science. Zenodo. https://doi.org/10.5281/zenodo.15351255
*A collection of articles on Citizen Science and AI is available through: https://theoryandpractice.citizenscienceassociation.org/collections/ai-and-citizen-science
Citizen Science and Open Science working group
Blog post author: Fotis Mystakopoulos (https://orcid.org/0000-0002-9354-3754)
Blog post contributors:
Étienne Serbe-Kamp (https://orcid.org/0000-0002-7097-6274)
Dilek Fraisl (https://orcid.org/0000-0001-7523-7967)
Karen Verstraelen (No ORCiD Found)
Blog post review:
Gefion Thuermer (https://orcid.org/0000-0001-7345-0000)
Loreta Tauginienė (https://orcid.org/0000-0002-3001-2200)