Citizen Science and AI - a fix or a feature?

Did you see this story in the press this month, or maybe you caught the link in the Telraam September newsletter? Citizen science has been in the news recently in Australia and around the world:

This caught my attention because it added the potential of AI (or Artificial Intelligence) to the analysis of the data, which is also something that will be coming to the Telraam Sensor 2 (S2) in the near future. There is an interesting differentiation, however, that separates the two approaches.

Much Citizen Science is carried out by the most motivated users who may share a goal or mission, but do not always have consistent ways of carrying this out, or methodologies. In the case of the Citizens of the Great Barrier Reef data gathering exercise, tens of thousands of individuals took their own images of the reef and its health. The ‘data’ generated would be amazingly useful, but at the same time varied enormously in terms of resolution, quality and accuracy of identification, timeliness, and crucial parameters such as geolocation/spatial information.

Making sense of this vast data is incredibly complex and time-consuming, so this is one way that AI can add value, which is to interpret the data faster & better than humans. It offers the researchers a companion tool to massively improve the productivity of their interpretation work.

However, this is interpretation AFTER the event, so the overall process will take a great deal of time to gather, analyse and report. In this way, AI is used to fix an issue with the existing data gathering process.

The way that Telraam plans on integrating AI is different.

With the launch of the Telraam S2, AI will be added to the data gathering process by including this capability with the sensor, and interpreting images as they are generated. What this means is that the data is of higher accuracy, consistency and timeliness. There’s no ‘fix’ needed since everyone is using the same device in more or less the same conditions - it is only the context that varies.

Any citizen science activity should be celebrated, but taking action locally relies on having reliable data as soon as possible, and this development will not only make the device easier to install for more people, but also provide better, faster and more accurate counts.

It will be interesting to see how AI adds to citizen science projects in the future. Do you know of any other examples?