Artificial Intelligence is rapidly becoming a leading tool in all aspects of research, consultancy and management. We see the value of AI in many areas, but are aware of the risks as well. We have for instance already seen some publications on wetlands in SE Asia being largely AI generated, as well as some data analysis underpinning restoration. The results are not always convincing and risk blurring lines on what is science-based management.
Therefore, we decided on a protocol for how we use AI, so people we work with know what they can expect from us.
We will use AI (Claude and others) in these ways:
- To streamline complex software development, e.g. our algorithms to determine canal water depth from satellite LiDAR data flight lines at different angles from canals have been optimized this way.
- For literature searches, e.g. to find the latest scientific publications on a certain topic, alongside our own searches – ensuring completeness.
- To search and summarize specific publications on specific topics, with targeted questions.
- To access ‘grey’ information on topics that have not been covered in scientific publications but are discussed in media and hidden reports (e.g. local flooding or fire history).
We will not use AI for:
- Full data analyses (only partial in a few cases, and supervised)
- Text writing (only streamlining of some individual paragraphs)
Bottom line: when working with us, you will know our products are well considered and based on our expertise and a broader body of scientific publications, enhanced by AI but not replaced by it.