Teaching & Learning
- Use of AI in assignments is governed by instructor discretion
- Students remain responsible for factual errors and false references in AI-assisted work
- If a student submits AI-generated assignment work without telling the instructor, it is treated as a breach of academic conduct and can be sanctioned; if AI use is allowed, it must be clearly referenced, including source material and prompts where applicable
If a student submits an assignment produced by an AI content generation tool without stating the tool use to the instructor, the student infringes good academic conduct in accordance with §16, 2 of the Study & Examination Rules of the MPP, MIA, MDS and §15, 2 of the Study & Examination Rules in the EMPA. This infringement will be sanctioned according to §17 of the Study & Examination Rules of the MPP, MIA, MDS and §16 of the Study & Examination Rules of the EMPA.
If instructors allow the use of AI tools for the preparation of assignments, their use needs to be clearly referenced in the text, even if the machine-generated text has been modified by the author.
Note that in all cases of using AI for presentation, source material and prompts (or other coding) should be clearly identified such that similar output could be replicated. Such details may be placed in footnotes, citations, or supplements as appropriate.
Students are responsible for factual errors and false references in their assignments, even if these have been provided by AI tools that were properly referenced. The assignment will be downgraded accordingly in such instances.
It is at the discretion of every instructor at the Hertie School to encourage or limit the use of AI tools in alignment with the learning objectives and examination requirements of their course.
- The guidance also directs instructors to design assignments and evaluation exercises to avoid AI access and reduce cheating risk where relevant skills are being assessed
- The university does not set a single blanket rule allowing AI in examinations; instead, instructors may encourage, limit, or entirely ban AI depending on course learning objectives and examination requirements
These guidelines outline the use of AI tools in the teaching, learning and student evaluation at the Hertie School.
It is at the discretion of every instructor at the Hertie School to encourage or limit the use of AI tools in alignment with the learning objectives and examination requirements of their course.
When skills potentially performed by using AI tools are the content of the course learning objectives and/or the examination requirements, the course instructor can decide to ban AI-tool use entirely. Instructors should plan their assignments and evaluation exercises in a way that avoids granting access to AI tools and limits the possibility for cheating.
- The university encourages sensible and responsible use of AI for learning and study support, including researching, summarizing, simplifying domain knowledge, and brainstorming questions
- Students are expected to critically assess AI outputs, combine AI search with traditional methods, maintain academic integrity, and avoid uploading protected, personal, confidential, or restricted data
AI content generation can provide valuable help for many tasks if used reasonably. As such, we want to enable our students to become acquainted with them.
The Hertie School encourages instructors to experiment with AI tools in teaching and learning to create innovative ways to acquire and organize knowledge and check the veracity of information AI-generated.
Research using AI tools: Using AI tools to quickly research, summarize or simplify key aspects of new domain knowledge, or brainstorm and explore research questions.
Critically assess the results suggested by AI: publication credibility, author authority, relevance to the topic
Combine AI search with traditional database searches and library systems to ensure comprehensiveness
Respect copyright, privacy and comply with GDPR. Do not upload protected or personal data
All Hertie School students have access to Hertie-AI via ai.hertie-school.org. The platform allows you to compare outputs from different large language models (LLMs), use commercial OpenAI ChatGPT, and collaborate with peers by creating shared “rooms” for prompts. Please note that prompts are processed outside Hertie School infrastructure. Do not enter confidential or restricted data. Results should always be critically assessed, as LLMs may produce inaccurate information.
- AI code-related use is left to instructor discretion
- The guidance explicitly states that programming courses may prohibit tools that translate natural language into code while allowing other AI-generated natural language support if it is properly referenced
It is at the discretion of every instructor at the Hertie School to encourage or limit the use of AI tools in alignment with the learning objectives and examination requirements of their course.
When the learning objectives focus on one specific skill, but not others proposed by AI tools, the instructor needs to clearly exclude the specific set that is not allowed in the class or in assignments.
For example, a course of programming language can exclude tools that translate natural language into the relevant programming language but allow for natural language generation to accompany the assignment, if it is properly referenced.
Research
- For master's theses and doctoral dissertations, students must declare independent authorship and disclose the sources and means used
- The provided sources do not set a specific university-wide rule on using generative AI to draft or edit research manuscripts, but they do require disclosure of used means in thesis/dissertation submissions and require AI-generated assignment text to be referenced when allowed
When submitting the Master's thesis, each student must make a written declaration that she/he worked on the thesis independently and did not use any other sources and means than those indicated.
The doctoral candidate shall present all forms of resources and aid and give assurance that the work was completed independently on that basis.
If instructors allow the use of AI tools for the preparation of assignments, their use needs to be clearly referenced in the text, even if the machine-generated text has been modified by the author.
- It also prohibits entering protected, personal, confidential, or restricted data into AI tools and notes that prompts may be processed outside Hertie infrastructure
- The university permits AI to support research-related searching and question refinement, but users must critically assess outputs for bias, outdated information, and inaccuracies
AI-powered search can quickly scan large amounts of data using Large Language Model technology to suggest context-aware results and help you refine your research questions. However, AI may miss context or suggest biased or outdated information.
Critically assess the results suggested by AI: publication credibility, author authority, relevance to the topic
Respect copyright, privacy and comply with GDPR. Do not upload protected or personal data
All Hertie School students have access to Hertie-AI via ai.hertie-school.org. The platform allows you to compare outputs from different large language models (LLMs), use commercial OpenAI ChatGPT, and collaborate with peers by creating shared “rooms” for prompts. Please note that prompts are processed outside Hertie School infrastructure. Do not enter confidential or restricted data. Results should always be critically assessed, as LLMs may produce inaccurate information.
- The sources do not explicitly mention AI-specific rules for grant proposals or IRB-style applications
- The university requires compliance with rules of good research practice and with legal and ethical frameworks in research
- Researchers must seek approvals and ethics statements where necessary, and the school has a research ethics review mechanism open to all members, including master students, for research involving ethical questions
The Hertie School, with the participation of Hertie School members, works to define rules of good research practice, ensuring that all Hertie School members are made aware of these guidelines and related policies and regulations, and requires all Hertie School members to comply with them with due regard for the type of research undertaken in the relevant subject area.
Researchers comply with rights and obligations, particularly those arising from legal requirements and contracts with third parties, and where necessary, seek approvals and ethics statements and present these when required.
The Hertie School has set up a research ethics mechanism to advise, review, and provide ethics statements. The research ethics review involves two stages: self-assessment and full review.
The research ethics review mechanism is open to all members of the Hertie School, including master students, whose research might involve ethical questions that would need to be checked at any point during their research.
Academic Integrity
- Disclosure of AI use is required when AI contributes to assignment preparation
- Students must state tool use to the instructor, clearly reference AI-generated content in the text, and identify source material and prompts or other coding so that similar output could be replicated; the library also instructs users to cite sources directly or indirectly
If a student submits an assignment produced by an AI content generation tool without stating the tool use to the instructor, the student infringes good academic conduct in accordance with §16, 2 of the Study & Examination Rules of the MPP, MIA, MDS and §15, 2 of the Study & Examination Rules in the EMPA.
If instructors allow the use of AI tools for the preparation of assignments, their use needs to be clearly referenced in the text, even if the machine-generated text has been modified by the author.
Note that in all cases of using AI for presentation, source material and prompts (or other coding) should be clearly identified such that similar output could be replicated. Such details may be placed in footnotes, citations, or supplements as appropriate.
Always cite sources directly (text) or indirectly (concepts, ideas, abstracts, summaries)
- Undisclosed AI-generated assignment submission is treated as a violation of academic conduct and is sanctionable under the school's study and examination rules
- The university also notes that instructors may use AI-generated text recognition tools for teaching or teaching preparation, but the sources do not define a formal AI detection policy
- Suspected breaches are referred by the lecturer to the Examination Committee, which may impose grade reduction, failure of the assignment or examination, or expulsion in severe or repeated cases
If a student submits an assignment produced by an AI content generation tool without stating the tool use to the instructor, the student infringes good academic conduct in accordance with §16, 2 of the Study & Examination Rules of the MPP, MIA, MDS and §15, 2 of the Study & Examination Rules in the EMPA. This infringement will be sanctioned according to §17 of the Study & Examination Rules of the MPP, MIA, MDS and §16 of the Study & Examination Rules of the EMPA.
In case of a suspected violation of the standards of good academic conduct, the lecturer shall submit the matter to the Examination Committee.
If the Examination Committee determines any violation of the standards of academic integrity, it shall recommend the following sanctions, taking account of the principle of proportionality:
a) grade reduction for the respective assignment,
b) failure of the respective assignment,
c) failure of the entire examination,
d) In case of particularly severe, repeated or multiple violations of the standards of academic integrity, the student shall be expelled from the Hertie School.
In addition, instructors may want to utilize AI tools for their own teaching or teaching preparation. Such tools can include an automated feedback coach to help students improve their writing, AI generated text recognition tools, test questions, sample answers, slide images, etc.
Institutional & Administrative
- Faculty are encouraged to experiment with AI in teaching and learning, and each instructor has authority to set course-level AI rules aligned with learning objectives and assessment requirements
- The guidance also states that instructors may use AI for teaching preparation, including feedback support, text recognition, test questions, sample answers, and slide images, and should address privacy issues with students
It is at the discretion of every instructor at the Hertie School to encourage or limit the use of AI tools in alignment with the learning objectives and examination requirements of their course.
In the Spring term 2023, the course policy on AI tools should be clearly discussed in class and posted in Moodle to guide students in the preparation of their assignments.
Reference to privacy: Instructors should sensitize students regarding privacy issues. Individuals using the tools remain responsible for data protection and for how they use personal information.
The Hertie School encourages instructors to experiment with AI tools in teaching and learning to create innovative ways to acquire and organize knowledge and check the veracity of information AI-generated.
In addition, instructors may want to utilize AI tools for their own teaching or teaching preparation. Such tools can include an automated feedback coach to help students improve their writing, AI generated text recognition tools, test questions, sample answers, slide images, etc.
- The university provides access to Hertie-AI for students and explicitly warns that prompts are processed outside school infrastructure
- Users must not enter confidential, restricted, protected, or personal data into AI tools, and broader network-use rules also prohibit unauthorized access, disclosure, or transmission of confidential or sensitive information without proper controls or authorization
Reference to privacy: Instructors should sensitize students regarding privacy issues. Individuals using the tools remain responsible for data protection and for how they use personal information.
Respect copyright, privacy and comply with GDPR. Do not upload protected or personal data
All Hertie School students have access to Hertie-AI via ai.hertie-school.org. The platform allows you to compare outputs from different large language models (LLMs), use commercial OpenAI ChatGPT, and collaborate with peers by creating shared “rooms” for prompts. Please note that prompts are processed outside Hertie School infrastructure. Do not enter confidential or restricted data.
Unauthorized reading/access of Hertie School administration information/material/ data
Misusing, disclosing without proper authorization, or altering of Hertie School administration information/material/ data
Transmission of any proprietary, confidential, or otherwise sensitive information without the proper controls or authorization
- The university has articulated an institutional approach that aims to enable sensible use of AI while limiting abuse, using a flexible model that leaves course-level implementation to instructors
- It has issued AI guidelines for teaching and learning, promotes responsible experimentation with AI, provides student access to Hertie-AI, and maintains a broader research ethics mechanism for institutional oversight of ethically sensitive research
The Hertie School believes it is crucial to make sensible use of such tools, and where possible to limit or prevent potential abuse.
As such, we want to enable our students to become acquainted with them. However, it is also clear that machine-generated work handed in as one’s own is a violation of academic integrity in the same way as plagiarism or other forms of authorship fraud. Hence, we opt for a flexible approach.
These guidelines outline the use of AI tools in the teaching, learning and student evaluation at the Hertie School. They seek to establish what constitutes sanctionable behavior and which type of use can be encouraged in the classroom, depending on the learning objectives and examination requirements set by the instructor.
The Hertie School encourages instructors to experiment with AI tools in teaching and learning to create innovative ways to acquire and organize knowledge and check the veracity of information AI-generated.
All Hertie School students have access to Hertie-AI via ai.hertie-school.org.
The Hertie School has set up a research ethics mechanism to advise, review, and provide ethics statements.
DocuMark: Responsible AI Use for Academic Integrity
Knowing your institution's AI policy is step one. DocuMark helps enforce it fairly by empowering universities to manage AI-generated content, prevent cheating, and support student writing through responsible AI use.
Common Questions About Hertie School (Berlin)'s AI Policies
Hertie School (Berlin) has defined AI policies in 12 of 12 categories, with an overall coverage score of 100%.
Disclosure of AI use is required when AI contributes to assignment preparation. Students must state tool use to the instructor, clearly reference AI-generated content in the text, and identify source material and prompts or other coding so that similar output could be replicated; the library also instructs users to cite sources directly or indirectly.
Undisclosed AI-generated assignment submission is treated as a violation of academic conduct and is sanctionable under the school's study and examination rules. Suspected breaches are referred by the lecturer to the Examination Committee, which may impose grade reduction, failure of the assignment or examination, or expulsion in severe or repeated cases. The university also notes that instructors may use AI-generated text recognition tools for teaching or teaching preparation, but the sources do not define a formal AI detection policy.
The university provides access to Hertie-AI for students and explicitly warns that prompts are processed outside school infrastructure. Users must not enter confidential, restricted, protected, or personal data into AI tools, and broader network-use rules also prohibit unauthorized access, disclosure, or transmission of confidential or sensitive information without proper controls or authorization.