Generative AI – Beyond Euphoria and Simple Solutions (2024)

Since programs such as ChatGPT and Dall-E have become available to the general public there has been intense discussion about the risks and opportunities of generative artificial Intelligence (AI). Due to their ability to create texts, images, and videos these programs can greatly benefit people’s everyday lives, but can also be misused to create deep fakes or propaganda. In addition, all forms of generative AI reflect the data used to train them and thus the objectives underpinning their development. Both aspects elude control by institutions and traditional norms. There are now some strategies to counteract the lack of transparency and objectivity (bias) of generative AI. However, the authors of the discussion paper, published today in English by the German National Academy of Sciences Leopoldina, warn against placing too much faith in these strategies. They take a realistic look at the possibilities and challenges regarding the development and application of generative AI.