AI-Chat
In the AI chat at HWR Berlin, a selection of different language models of text-generating artificial intelligence (AI) is available.
Use AI responsibly
Use AI at HWR Berlin responsibly: Our policy provides clear guidelines for ethical, legal, and data-protection-compliant use.
HWR Berlin Policy on the Use of AI
(in German)
In addition, follow the requirements set by instructors/examiners and the faculties regarding the use and labeling of AI.
The Use of Generative AI and Data Protection
(in German)
Drafting aids for declarations of Authorship
(in German)
In this five-part video series, central aspects of artificial intelligence are explained in an easy-to-understand way. The topics range from technical basics to societal questions, as well as practical applications in organizations, creativity, and research. The focus is not only on opportunities, but also on challenges: How can AI be used responsibly? What ethical, legal, and cultural questions do we need to consider? The accompanying article provides a compact overview, places the video content in context, and invites you to understand AI as a tool we can shape in the present time.
Basics of Artificial Intelligence
“Generative AI calculates likely content; it does not generate knowledge.”
Technical fundamentals (in German)
Was genau ist Künstliche Intelligenz – und wie funktioniert sie? In diesem Video erklären wir die technischen Grundlagen von KI und generativer KI leicht verständlich: von den zentralen Bausteinen wie Daten, Algorithmen und Rechenleistung bis hin zu maschinellem Lernen, neuronalen Netzen und Deep Learning. Anhand anschaulicher Beispiele zeigen wir, wie KI bereits heute in unserem Alltag wirkt – etwa in Sprachassistenten, Navigationssystemen oder Empfehlungstools. Außerdem werfen wir einen Blick auf die verschiedenen Arten von KI – von einfachen reaktiven Systemen bis hin zur theoretischen Vision einer selbstbewussten Superintelligenz. Das Video bietet einen kompakten Überblick über den aktuellen Stand der Technik und beleuchtet zugleich die Herausforderungen, die mit der weiteren Entwicklung leistungsfähiger KI-Systeme verbunden sind.
Societal reflection
(in German)
Create creative content
(in German)
Organization and automation
(in German)
Research and data analysis
(in German)
AI pilot phase
Since the introduction of HAWKI for the 2024/25 winter semester, members of HWR Berlin have had access to a generative AI tool that complies with the GDPR.
Alongside the one-year pilot phase with HAWKI, alternative solutions were sought to better meet evolving requirements.
The launch of the new AI chat service based on Open WebUI marks the beginning of a structured transition. Open WebUI replaces the existing HAWKI AI service, which is scheduled to be phased out by the 2026/27 winter semester.
Are personal data sent to OpenAI?
No. OpenAI only receives the prompts you enter. Neither user data nor IP addresses are shared because the HAWKI server handles the forwarding. The use is anonymous, which means storing chat histories on a user account is not possible.
Are data backed up, monitored, or analyzed?
No. There is no storage, monitoring, or analysis of data. Your inputs are not used to train the AI at OpenAI, which is ensured by specific settings in the API. However, for technical analysis of misuse and misappropriation, OpenAI will retain your anonymous requests on the OpenAI servers for up to 30 days.
Do costs arise from using it?
The HWR Berlin covers the costs. They are calculated based on the word count of your inputs plus the responses.
So asking short and precise questions saves money!
Are there rules for use?
Yes, the main rules are:
- Do not include personal data in your prompts.
- Do not include copyrighted material in your prompts.
- Use of HAWKI is permitted only within the scope of your studies or official duties.
The terms of use must be followed.
Are there recommendations for dealing with the results?
The answers from generative AI are largely correct, but they can also be outdated, wrong, or even fabricated. Users are responsible for checking the results for accuracy. Directly using the results is not recommended. Decisions should not be made solely on the basis of generative AI’s outputs. Learn about how generative AI works in general, as well as its advantages, disadvantages, and risks.
Will there be more language models?
The available portfolio of language models changes regularly because the provider releases new versions and removes outdated models. Once a model is discontinued by the provider, it is no longer available in our AI chat.
We commit to integrating new models into our system only after a thorough review with regard to data protection and data security.
FAQ
Frequently asked questions
Glossary
explained simply
Generative AI
An area of artificial intelligence that focuses on models capable of generating content—such as texts, images, music, or code—that resembles what humans might create.
Prompt
A prompt is an input or a starting text that you give to an AI to receive a response or a continuation. It serves as guidance for the AI so it understands the context or topic it should respond to.
Prompt-Engineering
Prompt engineering is the process of creating and optimizing input prompts for AI in order to get the desired answers or results. It involves carefully wording and adjusting the prompts to encourage the AI to provide as precise, relevant, and high-quality responses as possible. The goal of prompt engineering is to make the best use of the AI’s capabilities and to steer the output.
Finetuning
Fine-tuning in the context of AI—especially with text-generating models—means adjusting or further developing a pre-trained model using specific data so that it performs better for particular tasks or applications.
Machine learning (ML)
A method of AI that uses algorithms to learn from data and make predictions or decisions without being explicitly programmed.
Deep Learning
A subgroup of machine learning based on artificial neural networks with many layers (“deep”). Deep learning models are particularly effective at processing image, video, and speech data.
Neural network
A network of artificial neurons, inspired by the structure and function of the human brain, used to build and train ML models.
Reinforcement Learning (RL)
A type of machine learning in which an agent learns to act in an environment by optimizing its actions based on rewards.
Natural Language Processing (NLP)
An area of AI that deals with the interaction between computers and human (natural) language—especially how to teach computers to read, understand, and generate human language.
Text-to-Image Generation
The use of generative AI models to create visual content based on textual descriptions.
