The idea behind the avatar

For many museum visitors, seeing the works of Käthe Kollwitz gives rise to questions and thoughts that they would like to share and discuss with others. In order to fulfil these wishes and strengthen the connection to the younger generation, the idea of the Kollwitz avatar was born – a virtual contact person who picks up on emotions and thoughts and acts in an attentive and listening way.

Content preparation: What would Käthe say today?

The first step was to collect frequently asked questions from visitors – from Kollwitz’ everyday life to the impact and significance of her art, as well as her possible stance on current social issues. The question ‘What would you do today?’ was asked particularly often. – for example in the context of current wars or social injustice. In her works, Kollwitz portrayed people who are affected by poverty, have lost family members in war or live on the streets – themes that suggest a connection to the present day. The aim was to prepare the chatbot for difficult questions. Selected data on Kollwitz’ artworks, diary entries and her biography formed the database.

Technical Realisation of the Project

The decision was made in favour of a chat and voicebot when considering the use of new technologies, due to its ability to access extensive knowledge and answer a wide range of questions. Working closely with IT intouch in Bergisch Gladbach, selected content was prepared for the software and tested with various so-called “prompts”. The avatar was given a matching voice for this selected personality. In addition, a connection had to be developed that allowed the animated avatar, the chatbot and the user interface on the touchscreen to operate smoothly together.

The project uses innovative technologies such as Google’s text-to-speech service for translation into numerous languages or AI-generated responses from OpenAI. The artificial intelligence (AI) is not only interactive, but also capable of learning (after a manual testing process) and now responds to questions individually and in genuine speech.

Insights into design development with Alla Popp.

Design challenges and artistic realization

Certain aspects cannot be reproduced by the AI – such as the ‘Berlin joking tone’ described by Kollwitz’ son Hans or a slight lisp. The design of the avatar itself was quite a challenge. The idea of animating a photo of the artist was soon discarded as it seemed too realistic and therefore a little spooky. It was important to make it clear that this is not Käthe Kollwitz herself, but a digital creature operated by an AI-supported system.

The Kollwitz avatar was designed by Alla Popp, an artist specialising in digital media. The 3D figure modelled exclusively for the museum forms a synthesis between Käthe Kollwitz’ appearance, her art and modern digital aesthetics. Various aspects of Kollwitz’ physical features and her art were incorporated into the design – such as the closed form of her bronze sculptures, the fine hatching of her prints, her empathetic smile and her modest style of dress.

Public Reactions

Initial tests with visitors of different ages showed very positive reactions to the avatar figure. The participants reacted playfully and enthusiastically. For digital natives in particular, the Kollwitz avatar offers a contemporary approach based on today’s media usage. Professional audiences and the press expressed great interest. Colleagues from all over the world visit the avatar and ask numerous, sometimes surprisingly challenging questions.

Critical perspectives: Sustainability and dependencies

Although the Avatar project is using innovative technologies, it also raises important questions regarding sustainability and long-term costs. The use of AI-supported systems requires continuous maintenance, updates and server costs. There is also a dependency on US companies such as OpenAI and Google, whose services are essential for the functioning of the avatar and may incur significant costs. This dependence can involve risks, for example in the case of changes to the terms of use, price increases or technical restrictions that jeopardise the independence and long-term stability of the project. In addition, the ecological balance of such digital projects is quite questionable, as the energy consumption for servers and data centres is not insignificant. Overall, while the project shows how digital innovations can enrich the museum experience, these aspects should be considered more in future projects to ensure the sustainable and responsible use of such technologies.

Uncertain future

The original plan was to continue the avatar project after the reopening of the Kollwitz exhibition and to permanently integrate the avatar into the educational work. This plan can be maintained until the end of 2025. However, it is uncertain whether it will be possible to use the avatar beyond this date, as long-term funding has not yet been secured.

Käthe Kollwitz died exactly 80 years ago – she is no longer able to comment on current events. Nevertheless, an attempt has been made to make her speak again with the help of innovative technologies. Her diaries contain statements that reveal a fundamental attitude and serve as the basis for the chatbot’s answers. The chatbot’s language system does not respond directly, but makes traits of her character clear. The answers that the AI generates based on the content provided are only partly under human control. The project therefore remains an experiment: for each question, the AI recombines the available data to find a suitable answer – always in the balance between historical authenticity and technical interpretation.