New study: citizen science projects are already using AI to cut training barriers — but legal/ethical guidance is "urgently required"
This covers a coding tool or code-capability update — useful for developers assessing workflow changes and reusable value.
A new study in PLOS ONE reveals that citizen science projects are leveraging AI to analyze volunteer-uploaded photos, reducing training needs and expanding participation.…
Co-designed study just out in PLOS ONE looking at what's actually working and not working in citizen science. A few AI-specific findings worth sharing here:
Where it's helping: Projects like the Great Reef Census and NOBURN use AI to analyse photos uploaded by volunteers, cutting reliance on expert training and reducing manual processing/human error. This is opening participation to people who'd otherwise be excluded by steep training requirements.
Where it's not keeping up: The ethical use and environmental impact of these tools remains largely unclear, and our respondents flagged that legal/ethical guidance from governments and institutions is lagging well behind adoption. We're recommending that any AI use in citizen science be transparently reported — including underlying code and training data — not just the outputs.
Broader tension: As AI lowers the skill barrier to contribute, the paper argues recognition and payment structures haven't caught up. If AI increasingly does the "labour" citizen scientists used to be trained for, questions about what deserves compensation and credit shift too.
Open access, all data on OSF, STARDIT report about the article Too:
https://doi.org/10.1371/journal.pone.0331161
Curious what this sub thinks about the governance gap here — feels like a pattern that shows up in a lot of participatory/public-facing AI use, not just citizen science.