Wuppertal Institute Publishes its AI Policy

Whether it's analysing large amounts of data in the blink of an eye, conducting comprehensive literature reviews or exploring ideas for new research topics using a machine’s ability to scan the entire internet, artificial intelligence (AI) has made all of this possible – and much more besides. Many people are increasingly feeling the impact and potential of large language models (LLMs for short) in the workplace. The Wuppertal Institute is no different and is systematically exploring both the current and future opportunities presented by AI in everything from research and administration, all the way through to communications.

However, the use of AI also comes with risks. By their nature, LLMs are highly dependent on the data on which they are trained. If care is not taken to ensure high quality and balanced training data, an LLM's output can be easily misinterpreted and may, for example, reinforce existing biases. AI models may also make mistakes, and it is essential that their output is always thoroughly reviewed – the human factor is irreplaceable when using AI. In addition, the use of AI requires a substantial amount of energy and, if used unnecessarily, it can cause harm to the climate and the environment. And when it comes to communication, a lack of transparency in the use of AI can damage trust between people – transparency is therefore of paramount importance, particularly in the scientific community.

In order to seize the biggest opportunities afforded by AI while simultaneously minimising the risks associated with its use, the Wuppertal Institute has developed its own AI policy, which came into effect at the end of March 2026. It defines guidelines for the considered, transparent and responsible use of AI in the Institute's day-to-day work. In the interest of transparency, this policy has been published on the Institute's website – thereby also demonstrating externally its conscious approach to the potential and challenges of AI.

The AI policy is to be continuously updated in line with current developments. It is deliberately aligned with other key AI guidelines for the scientific community, such as those of the German Research Foundation.

The Wuppertal Institute's AI policy can be accessed and downloaded via the link below.