We work intensively on current legal questions around artificial intelligence and large language models: who owns AI-generated content, who is liable for errors, and how data protection and transparency can be ensured. Our experience, which we deepened further in a state-funded research project, helps you assess these questions and find practical solutions.
Liability provisions, information obligations, allocation of obligations in standard terms and individually negotiated agreements. We write clauses that address AI-specific risks cleanly — so they don't have to be fought out in court first.
From the EU AI Act (KI-VO) and the Data Act through to copyright and data protection. We classify AI systems by risk level, assess conformity, and track the regulation's staggered application: prohibited practices since February 2025, GPAI obligations since August 2025, transparency obligations under Art. 50 since August 2026. The high-risk obligations apply under Regulation (EU) 2026/1744 (Digital Omnibus on AI) from December 2027 (Annex III) and August 2028 (Annex I), respectively.
Keynotes, talks, webinars, employee training, and publications. These include our five-part webinar series “Recht trifft KI” (“Law Meets AI”) and our briefing “KI und Digitalisierung” (“AI and Digitalization”), both with Deutscher AnwaltVerlag, as well as regular talks and workshops on AI-generated code, copyright, and compliance.
Legal support for building your own AI applications: software development contracts, licensing of the output, data processing agreements, NDAs, and the transparency obligations under Art. 50 KI-VO. We know the pitfalls, because we have walked this path ourselves.
In a state-funded research project, we researched legal questions that cannot be answered from textbooks alone.
The EU AI Act separates entry into force from application: it has been in force since August 2024 and becomes applicable in stages. Regulation (EU) 2026/1744 (Digital Omnibus on AI) reordered that staggered application at the end of July 2026, six days before the original high-risk deadline. Since the KI-MIG (Germany's AI market surveillance and innovation promotion act) took effect, the competent market surveillance authority in Germany has been the Bundesnetzagentur (Federal Network Agency); the existing sectoral supervisory authorities retain their remits.
For connected products with an AI component, the Cyber Resilience Act applies in addition. For high-risk AI systems, its market surveillance lies under Art. 52(14) CRA with the authority designated for the EU AI Act — we cover this interface under Cybersecurity & data economy.
The EU AI Act, Regulation (EU) 2024/1689, enters into force. Application follows in stages under Art. 113.
Prohibited practices (Art. 5) and the AI literacy obligation (Art. 4) apply.
The obligations for GPAI models and the governance provisions apply.
The Digital Omnibus enters into force. It reorders the high-risk timetable and narrows the high-risk definition.
The KI-MIG enters into force. The Bundesnetzagentur becomes the central market surveillance authority, point of contact, and complaints body.
General application of the EU AI Act. The transparency obligations of Art. 50 apply: disclosure when interacting with AI systems, labeling of deepfakes, machine-readable marking of synthetic content for systems newly placed on the market.
End of the grace period for machine-readable marking for systems placed on the market before August 2, 2026. In addition, two new prohibitions under Art. 5 apply.
The obligations for high-risk AI systems under Annex III apply (Chapter III, Sections 1 to 3).
The obligations for high-risk AI systems under Annex I apply, that is, for AI in regulated products such as machinery or medical devices.
The practitioner's handbook “Generative KI in der Rechtsberatung — Erkenntnisse zum Einsatz von Large Language Models” (Generative AI in Legal Advice — Insights on the Use of Large Language Models; in German) offers sound insights into the practical use of LLMs in legal practice. From the editors' own benchmark findings to day-to-day legal advice. Alongside an accessible look at the history and technology, the focus is above all on real-world experience with the use of AI in everyday legal work.
What opportunities does the use of AI offer legal departments and law firms? Which risks and legal pitfalls need to be considered? The three editors take you into a highly topical field between innovation, regulation, and reality: Dr. Sophie Garling, Annika Niemann and Yvonne Roßmann.
Numerous examples, valuable tips on using prompts in legal practice, and an overview of current points of dispute and case law round out the work. A hands-on book suited to lawyers, law firms, anyone interested in legal tech, and decision-makers alike.
Five-part webinar series with Deutscher AnwaltVerlag (“Law Meets AI”; held in German) on legal questions surrounding generative AI in legal practice. Held twice already; the next edition begins on November 6, 2026. Practice-oriented, with case studies from client work, a prompting workshop, and certification as a “GenAI Legal Expert.”
Regular talks on how AI-generated source code behaves under licensing law and which copyright risks arise in practice when FOSS components are carried over, in paraphrased form, into proprietary code by LLM tools.
Beyond that, we speak on the general copyright questions surrounding generative AI: from training on copyright-protected works, to the protectability of AI-generated outputs, to license design in human-machine collaborations.
Practitioner's handbook from JUN Legal GmbH, edited by Dr. Sophie Garling, Annika Niemann and Yvonne Roßmann. More in the highlight above.
Our briefing (“Infobrief”; “AI and Digitalization: On the Future of Law,” in German) with Deutscher AnwaltVerlag appears every two months and bundles the most pressing questions around AI and digitalization in the law — from generative AI as a strategic partner in the law firm, through prompting and large language models, to copyright, data protection, and further digitalization topics of legal practice. Prepared with practice in mind for legally sound counseling, supplemented by special issues on particularly current developments.
Our formats build on the participants' prior knowledge: from the basics of LLMs and how they work, through hands-on guidance on tool use, to deep dives into AI compliance, contract drafting, and data protection. Dates on request through our office.
When AI systems are introduced in a company, works councils face two challenges at once — new codetermination matters and a significant need for training in order to assess AI systems competently in the first place.
We support both: codetermination under the BetrVG (German Works Constitution Act) when AI systems are introduced, as well as our engagement as external AI experts for the council — supplemented by council-specific trainings on how AI works, its risks, and the practical assessment of specific AI applications in the workplace, also in combination with our data protection trainings for works councils.
