Remote Sensing Scientists and Technologists

However, not everything is automated. We found few examples of AI choosing new hardware or doing fieldwork – these still need human decision-making. Developing new correction algorithms or hardware recommendations requires expert judgment, and putting boots on the ground for field surveys is obviously a human job.

In short, AI is already augmenting many core tasks (especially the heavy data crunching), but scientists still guide the process and handle any tricky issues or quality control.

On the other hand, remote sensing data can be complex. Good AI models often need lots of labeled examples (for instance, human-curated maps) to learn from, and experts say this data can be hard to get ^[2]. Also, research shows that an AI model trained in one region may not work well in another without retraining ^[1].

This means scientists must spend effort to train and check the AI, which can slow adoption. In many agencies, skilled analysts and scientists are still required to interpret results and ensure accuracy.

Overall, experts emphasize that AI will be a helpful tool, not a replacement. It can handle routine analysis and let people focus on interpretation and problem solving ^{[3] [2]}. As one NASA engineer said, these tools add to our “data-analysis toolbox” ^[3].

While it brings challenges, the hope is that AI will free remote sensing professionals from boring data work and let them use their human skills—like creativity and field knowledge—for the parts machines can’t do.