Potential of Artificial Intelligence for Citizen and Library Science

One would be hard pressed to avoid hearing about artificial intelligence (AI) in the news. It’s everywhere: from AlphaGo, the new world champion of the board game Go, to Mark Zuckerberg’s announcement on framing AI as the core technology of Facebook. With all of this buzz, what actually is artificial intelligence?

Artificial Intelligence Versus Machine Learning Versus Deep Learning

AI is the concept of machines that are able to perform tasks that would normally require human intelligence. Machine learning is a technique used to implement AI. This is done by “training” a computer model using large data sets. For instance, if I show a machine-learning model one million images of a turtle, that model would effectively be able to identify turtles. Deep learning is another component of AI that is inspired by the human brain. Neural Networks (NN’s) are computer networks that follow the same structure of a human brain to make decisions.

Conceptions and capabilities of AI can be wide-ranging, and while we’re nowhere near a conscious robot apocalypse, the technology is showing a lot of promise. AI can work impressively well with narrow tasks, such as image recognition and speech recognition. Apple’s Face ID and the Amazon Alexa, respectively, are two pertinent examples. AI algorithms can spot fraud, navigate vehicles, and recommend movies for you to watch.

The Limits to Small Datasets

All machine-learning or AI projects are constrained by the amount of data available. Because these models are so computationally expensive, they need thousands or millions of data points to create a working model. The biggest bottleneck for the process involves labeling data, which is identifying data so that a computer can learn from it. Labeling large-scale data is nearly impossible due to the time-consuming process of doing it manually. Impossible, that is, until the advent of citizen science.

How Citizen Science Can Help

Citizen science is the collection and analysis of scientific data conducted by members of the general public. While the term was originally coined in the late 1970s, it has recently gained much greater popularity with the internet and through organizations such as the Citizen Science Alliance. In 2008, the Citizen Science Alliance launched Zooniverse, a web platform used to connect crowdsourced scientific research. Zooniverse works by allowing users to sign up and participate in a variety of projects. Through specific instructions and a clear user-experience, individuals can classify images or organize digital research projects. As of 2018, Zooniverse has over 1.7 million registered users who have together submitted nearly 400 million classifications.

AI for Science

One of the most promising capabilities of AI is for scientific research. Now, with the image classifying power of Zooniverse, hundreds of AI implementations are now available to be experimented upon. The University of Michigan just completed a Zooniverse project called Michigan ZoomIN where participants identified animals and humans in images captured throughout the state of Michigan. This data will allow us at the Shapiro Design Lab to use image classification neural networks to find patterns in the data, assisting Ecologists researching mammal and carnivore interactions across the state.

We’re excited to be at the forefront of citizen science research and its further the impact of artificial intelligence on scientific discovery. Stay in touch with the Shapiro Design Lab to follow our journey!


Zevin, M et al. “Gravity Spy: integrating advanced LIGO detector characterization, machine learning, and citizen science” Classical and quantum gravity vol. 34,No 6 (2017): 064003.

Keshavan, Anisha, Jason Yeatman, and Ariel Rokem. "Combining citizen science and deep learning to amplify expertise in neuroimaging." bioRxiv (2018): 363382.

Bonney, Rick, et al. "Citizen science: a developing tool for expanding science knowledge and scientific literacy." BioScience 59.11 (2009): 977-984.

McClelland, Calum. “The Difference Between Artificial Intelligence, Machine Learning, and Deep Learning.” Medium, IoT For All, 4 Dec. 2017.

Russell, Stuart J., and Peter Norvig. Artificial intelligence: a modern approach. Malaysia; Pearson Education Limited,, 2016.