AI - An Orientation

Aritificial intelligence in research, teaching, appreciation and advancement of the arts, and design

Although generative artificial intelligence (AI) has increasingly found its way into our everyday lives in recent years and decades, e.g. voice control of electronic devices, the release of ChatGPT was a milestone. AI is used in different areas such as robotics, autonomous driving, medical diagnostics, voice recognition, text, image and film design and many more.

Artificial intelligence offers tremendous possibilities for research and teaching, but it also poses great challenges to everyone involved. At the University of Arts Linz, the different forms of artistic, creative and scientific exploration are closely linked in theory and practice. They are understood as an interrelated discourse that can be enriched by AI. Artificial intelligence itself can become the subject of discussion, e.g. when we critically explore the impact of AI on art, society, politics, the environment or the economy.

This text aims to provide orientation regarding the meaningful and creativity-promoting integration of AI in research and teaching.

AI in artistic, creative and scientific practice

The utilisation of artificial intelligence in artistic and scientific practice offers new opportunities but also challenges regarding ethical, copyright and methodological questions.

For instance, in many cases we do not know what data was used to train AI applications or the models they are based on. It might come from material protected by copyright law, material whose proprietor has not agreed to its utilisation. Inhumane content may have been used and the working conditions of the workers involved in the process may be dubious. Moreover, results may be sexist, racist, reinforce prejudice or glorify violence. Artists and scientists should be aware of the ethical implications of their actions and use AI in a manner corresponding to their respective discipline. This includes considering and avoiding potential negative effects on society and the environment. Also, results of AI-based analyses should always be questioned critically and reviewed using other methods, hence avoiding distortions and misinterpretations.

In the training of AI models, we should primarily use data that is either freely available, licensed or released by the proprietor. Open-source databases and creative-commons-licensed content can form a legally secure basis. However, artistic approaches that address authorship, e.g. as a continuation of appropriation art, can be a special case.

Artificial intelligence sees an increasing integration into systems and work processes, making it difficult to distinguish between what AI has generated and what creative or intellectual work has been performed by humans. The deliberate use of AI should be documented as transparently as possible.

An experimental and critical-reflective approach to artificial intelligenc

The application of and reflection on artificial intelligence expands the spectrum of expression and discourse. The University of Arts Linz enables and promotes free experimenting with AI technology in a protected and supportive environment. Students and researchers have the opportunity to test AI methods in creative ways and further develop individual expertise and competence in working with current and future technologies.
At the University of Arts Linz, critical analysis of social, ecological, technological and cultural developments is of central importance. AI is not only the subject of critical research, it is also used as a tool for analysing methods and processes. The use of AI entails new phenomena and risks, such as black box problems, data bias, marginalisation, neo-colonialist structures in data generation, deepfake technology etc. The University of Arts Linz promotes artistic and theoretical exploration that addresses these problematic aspects of AI, reflecting on their effects on artistic, ecological, political or social processes.

The University of Arts as a centre of teaching, research, appreciation and advancement of the arts, and design

As places of knowledge production and communication, universities play a central role in society. They have long been responsible for archiving knowledge in the form of libraries and databases. Art universities, connecting art and design to science, play a special role among universities and thus face special challenges regarding AI.
The creation of an archive serving as a basis for artificial intelligence aims at expanding this role. The university could act as a central place of access to large, high-quality datasets required for the development of AI applications. Austrian universities have therefore agreed on a joint AI project: ‘Academic AI Services’ aims to establish a secure environment for testing and using (generative) AI technology, taking into account data privacy and information security, and to make this technology accessible to research, appreciation and advancement of the arts, teaching and administration. AI databases at universities could be available to researchers, teachers and students of various disciplines, facilitating interdisciplinary exchange and the critical analysis of interactions between art, design, technology, culture and society. The establishment of such archives allows the University of Arts Linz to promote an open and fair model for database access, benefitting both large and small institutions as well as independent researchers and artists.

AI as administrative support in research, teaching, appreciation and advancement of the arts, and design

The University of Arts permits and promotes the use of technology-based software, e.g. for analysing and structuring data in application phases or research project reports. However, in order to enable a well-founded integration of AI in research and practice, a precise tracking of sources must be guaranteed. In the interest of transparency, it is necessary to record the use of software in the compilation of written documents.

What to consider when using AI systems

Handling data

Data is a key resource of the digital age. However, it only forms a basis for further processing. Hence, not data itself is important, but also how and where it was generated, secured, processed, used and disseminated. The General Data Protection Regulation must also be observed regarding the use of AI systems. Personal data, for instance, must not be processed in AI systems under any circumstance, as this would constitute a violation of the General Data Protection Regulation.

Keeping an eye on data sovereignty

Data sovereignty refers to the right of an individual or an organisation to retain control over their own data and to determine how it is used. It includes the security, privacy and confidentiality of data and plays a crucial role when it comes to handling personal information in the digital world. Hence, a careful approach to data-handling in the use AI systems is very important, as the free versions of most systems use the data entered to train their AI models. It is recommended to read the terms of use before using an AI system or model.

Use of AI tools for scientific work

Generally, the use of AI tools for scientific work is acceptable, but each use must be documented in detail, i.e. the tool itself, its use and the scope of its use must be documented.

Furthermore, Austrian copyright law stipulates that users of AI output do not acquire any copyright to the output. AI-generated content, especially regarding theses, is not to be assessed as an independent achievement. Regardless of this, the use of copyright-protected data by machine learning systems needs a critical approach, i.e. uploading content for which the user does not own the copyright and which can be further machine-processed is not permitted.