Change log

Technische Universität Berlin (TU Berlin)

Source-backed change history with no release-to-release policy diff rows recorded yet; current claims, official sources, review state, and freshness remain visible across 0 public release records.

Review: Agent reviewedLast checkedMay 14, 2026Last changedMay 14, 2026Claims8Sources6Release records0

University page Public JSON Monthly report University index

Technische Universität Berlin (TU Berlin) currently has 8 source-backed claim records and 6 official source attributions. Latest tracked changed date: May 14, 2026. No tracker diff rows are recorded in the latest public release.

This page combines all public release diffs for Technische Universität Berlin (TU Berlin). Individual release snapshots remain available from their release-specific URLs.

No release-to-release policy diff rows are recorded for this university yet. The page still tracks current source-backed claims, official source attributions, review state, source freshness, and public JSON for discovery and citation.

This tracker is not legal advice, not academic integrity advice, and not an official university statement unless a linked source is the university's own official page.

Newly extracted claims are tracker additions and are not necessarily newly published by the university. Source snapshot changes show hash changes for the same source URL and are not by themselves policy changes.

Policy text0Newly extracted0Evidence0Source snapshots0Source text0Source added0Source removed0

11 # Technische Universität Berlin (TU Berlin) AI policy record
2+privacy: TU Berlin's living AI guidelines say personal data may be entered into generative AI tools only when the operator neither makes the data accessible to third parties nor uses it as training data; confidential information, sensitive research data, and internal documents are not permitted inputs.
3+Evidence (de, acaac1a5c670): Jedoch dürfen personenbezogene Daten nur in generative KI-Tools eingegeben werden, wenn die Softwarebetreiber diese Daten weder Dritten zugänglich machen noch als Trainingsdaten nutzen. Über den Datenschutz hinaus ist auch die Eingabe vertraulicher Informationen, sensibler Forschungsdaten oder interner Dokumente nicht zulässig.
4+academic_integrity: TU Berlin examination guidance states a general documentation duty when AI is used for assignments: students should identify the tools, extent, purpose, marked AI content, prompts, and their own contribution.
5+Evidence (de, 37ee496795d6): Generell gilt eine Pflicht zur Dokumentation, wenn Sie KI zur Bearbeitung von Aufgaben verwenden: Geben Sie stets an, welche Tools Sie genutzt haben, in welchem Umfang und zu welchem Zweck ... Falls Sie KI-Inhalte ... verwenden, kennzeichnen Sie diese und benennen Sie die verwendeten Prompts.
6+teaching: TU Berlin's living AI guidelines say fully automated assessment of study or examination performance is not permissible; AI output may not be adopted one-to-one and digital tools such as AI may serve only as aids.
7+Evidence (de, acaac1a5c670): Eine vollständig automatisierte Bewertung von Studien- oder Prüfungsleistungen ist nicht zulässig (basierend auf Art. 22 DSGVO). Das heißt, der Output darf nicht eins zu eins übernommen werden, sondern digitale Tools wie KI dürfen nur als Hilfsmittel dienen.
8+academic_integrity: TU Berlin examination guidance tells students they remain fully responsible for all AI-generated content; errors or inaccuracies in AI output count as the student's errors, and AI may support work but may not take over the actual work.
9+Evidence (de, 37ee496795d6): Für alle durch KI generierten Inhalte tragen Sie die volle Verantwortung. Fehler oder Ungenauigkeiten in KI-Ergebnissen gelten als Ihre Fehler und nicht die des Tools. KI kann Sie unterstützen – etwa bei Ideenfindung, Recherche oder Strukturierung – darf aber nicht die eigentliche Arbeit übernehmen.
10+academic_integrity: For final and seminar papers, TU Berlin examination guidance says AI-supported passages must be marked, except spelling checks and stylistic improvements, and that use of AI tools must be discussed with the supervisor.
11+Evidence (de, 37ee496795d6): Alle durch KI unterstützten Passagen müssen kenntlich gemacht sein (z. B. durch Fußnote) – Ausgenommen sind Rechtschreibprüfungen und stilistische Verbesserungen. ... In jedem Fall müssen Sie die Verwendung von KI-Tools mit Ihrer:m Betreuer:in absprechen!
12+research: TU Berlin's living AI guidelines say that if AI results are adopted in scientific work, they must be checked for content, revised, and labeled; uses without relevance to learning or epistemic work, such as grammar, style, spelling, or translation, usually do not need to be declared.
13+Evidence (de, acaac1a5c670): Wenn jedoch KI-Ergebnisse übernommen werden sollen, müssen diese inhaltlich geprüft, überarbeitet und gekennzeichnet werden. ... KI-Nutzungen ohne Relevanz ... müssen meistens nicht angegeben werden („beispielsweise Grammatik-, Stil-, Rechtschreibprüfung, Übersetzungsprogramme“ ...).
14+ai_tool_treatment: TU Berlin's ChatGPT Edu pilot page describes scoped institutional access: the pilot began with Faculty VII, lists EU-server use, TUB-account/pseudo-ID login, and research/teaching/administration use, and says the licenses were extended with all faculties participating and 790 licenses total.
15+Evidence (de, e0ae1d4f6cb0): Datenschutzkonforme Nutzung auf Servern innerhalb der EU; Sichere Anmeldung mit TUB-Account und Pseudo-ID; Nutzung in Forschung, Lehre & Verwaltung erlaubt ... Die Lizenzen werden um ein weiteres Jahr verlängert ... alle anderen Fakultäten der TU Berlin beteiligt und die Gesamtzahl der Lizenzen auf 790 erhöht.
16+privacy: TU Berlin's data-protection blog recommends limiting personal data in ChatGPT, not sharing confidential information, avoiding personal data especially about third parties, and disabling chat history because otherwise communication is stored long term and may feed training data depending on settings.
17+Evidence (de, a2df06661636): Geben Sie nur wenige persönliche Daten preis ... Vertrauliche Informationen sollten nicht über chatGPT geteilt werden. ... sollten keine personenbezogenen Daten (insbesondere Dritter!) an chatGPT übermittelt werden. ... Chathistorie deaktiviert werden, denn andernfalls wird die Kommunikation langfristig gespeichert und fließt je nach gewählten Einstellungen als Trainingsdaten mit ein.

Release history

0 public release diffs

Claim changes

8 claim records

privacy

TU Berlin's living AI guidelines say personal data may be entered into generative AI tools only when the operator neither makes the data accessible to third parties nor uses it as training data; confidential information, sensitive research data, and internal documents are not permitted inputs.

Review: Agent reviewedConfidence96%Evidence1Languagesde

academic_integrity

TU Berlin examination guidance states a general documentation duty when AI is used for assignments: students should identify the tools, extent, purpose, marked AI content, prompts, and their own contribution.

Review: Agent reviewedConfidence95%Evidence1Languagesde

teaching

TU Berlin's living AI guidelines say fully automated assessment of study or examination performance is not permissible; AI output may not be adopted one-to-one and digital tools such as AI may serve only as aids.

Review: Agent reviewedConfidence95%Evidence1Languagesde

academic_integrity

TU Berlin examination guidance tells students they remain fully responsible for all AI-generated content; errors or inaccuracies in AI output count as the student's errors, and AI may support work but may not take over the actual work.

Review: Agent reviewedConfidence94%Evidence1Languagesde

academic_integrity

For final and seminar papers, TU Berlin examination guidance says AI-supported passages must be marked, except spelling checks and stylistic improvements, and that use of AI tools must be discussed with the supervisor.

Review: Agent reviewedConfidence94%Evidence1Languagesde

research

TU Berlin's living AI guidelines say that if AI results are adopted in scientific work, they must be checked for content, revised, and labeled; uses without relevance to learning or epistemic work, such as grammar, style, spelling, or translation, usually do not need to be declared.

Review: Agent reviewedConfidence93%Evidence1Languagesde

ai_tool_treatment

TU Berlin's ChatGPT Edu pilot page describes scoped institutional access: the pilot began with Faculty VII, lists EU-server use, TUB-account/pseudo-ID login, and research/teaching/administration use, and says the licenses were extended with all faculties participating and 790 licenses total.

Review: Agent reviewedConfidence90%Evidence1Languagesde

privacy

TU Berlin's data-protection blog recommends limiting personal data in ChatGPT, not sharing confidential information, avoiding personal data especially about third parties, and disabling chat history because otherwise communication is stored long term and may feed training data depending on settings.

Review: Agent reviewedConfidence90%Evidence1Languagesde