New AI Lab develops intelligent assistants for research and teaching

A new AI Lab at the Faculty of Natural Sciences will, in the coming years, serve as a driving force for the development and implementation of artificial intelligence in research and teaching. The lab is a strategic initiative designed to ensure that the university not only keeps pace with developments but places itself at the very front when it comes to responsible and targeted use of AI.

The AI Lab has been established as a 2.5‑year pilot project and is funded with DKK 12.4 million from the research reserve. The initiative stems from a strategic decision to invest in the university’s own AI solutions and infrastructure rather than relying on external tools. In addition to funding two AI specialists, the project has also included the purchase of two high‑performance AI servers for local execution of AI software at AU. The servers were acquired just before the significant rise in hardware prices triggered by increased global demand.

“To maintain a leading role and be the beacon we strive to be in a time marked by geopolitical uncertainty, continuous access to AI tools is essential,” says Poul Nissen, Vice‑Dean for Research at the Faculty of Natural Sciences.

The AI Lab targets educators, researchers, and students, focusing on two main areas: AI assistants (chatbots based on language models) and the development of AI tools that support existing research at the faculty.

Within AI assistants, the lab develops targeted language models tailored to specific academic needs of the researcher, educator, or student.

If you are an educator, this means that, in collaboration with the AI Lab, you can configure language models for your subject based on textbooks and course materials uploaded and used internally within the university’s systems. This ensures compliance with copyright and data security.

If you are a student, the AI assistant may feel like having access to your instructor around the clock. The AI help can be limited to individual classes and serve as an academic sparring partner that answers questions based on the syllabus.

The AI Lab does not only focus on the development and application of language models but works closely with research, with particular emphasis on generative and analytical AI. The goal is to integrate new AI algorithms from, for example, computer science with advanced datasets and complex problems in experimental research environments to enable new scientific breakthroughs.

The development of AI tools takes place through systematic investigations of application possibilities, as well as the scientific and methodological benefits of using AI across the faculty’s research domains. Projects are selected based on academic assessment of researcher inquiries and are carried out in interdisciplinary collaboration between researchers and AI developers.

One of the first concrete products from the AI Lab is a language model that supports researchers with research data management. Instead of completing complex forms without prior knowledge, researchers can input their project descriptions and receive step‑by‑step assistance in preparing data management plans.

Since 2020, researchers have been required to report data to the Danish National Archives, but the requirements can appear complex and daunting. Here, an AI assistant can also support researchers.

“We can free up an enormous amount of researchers’ time by reducing the bureaucratic burden they encounter daily, using role‑specific AI assistants for less interesting tasks. For example, preparing a data management plan is something researchers must do for each research project, but it is also time‑consuming. By letting AI assist the researcher in preparing a plan, we can combat a significant administrative burden and allow researchers to spend more time on their core tasks. However, it is always important to check and adjust what the AI assistants deliver,” explains Senior Consultant Jesper Lykkegaard Karlsen, one of the founders and advisors at the AI Lab.

In the long term, the same type of language model can also be used in connection with, for example, grant applications and much more.

Research groups will also be able to obtain their own AI assistants through the AI Lab, trained on relevant literature and data within the researcher’s field. The AI assistant can help the researcher with everything from coding and literature searches to quality assurance of scientific articles. For example, the researcher can ask whether a given problem has already been solved and by whom.

The AI Lab is intended to be an open environment in close interaction with researchers themselves. Since researchers are the users of the language models, their assessment of quality, relevance, and validity plays a central role in ongoing development.

The lab will be a significant contributor to the faculty’s new NAIS AI network. At the same time, there will be a strategic focus on outreach and knowledge sharing within the Faculty of Natural Sciences and across faculties.

Several faculties have already expressed interest in participating. The AI Lab will also collaborate with the Center for Humanities Computing (ARTS), focusing on exchanging experiences in a rapidly evolving field.

“The field is moving incredibly fast. Therefore, it is crucial that we do not lock ourselves into a single solution but continuously adjust and develop,” says Jesper Lykkegaard Karlsen.

The AI Lab is a pilot project running from 1 February 2026 to mid‑2028.

This article is translated with help from Copilot.