
Iveta Yuskeselieva
Technology Legal Counsel
Difficult legal questions,
answered.
More about meTechnology moves faster than contracts and regulations meant to govern it. I work in the gap between them: decoding what your obligations actually require, what your software actually does, so the space between the two stops being a risk you can’t see.
The questions that reach me tend to sit where the code, the contract and the regulation each say something slightly different. Finding where they disagree, before it becomes a problem, is the work I chose, and the work I’m genuinely good at.
What brings you here?
Featured series
Beyond the
source code
Exploring the legal questions that software licensing has yet to answer.
View all in seriesPhotoCraft: the legal questions an AI-assisted rewrite leaves open
PhotoCraft’s AI-assisted Photoshop alternative raises questions about Adobe’s licence terms, EU research rights, patents and the lawful use of open-source code
OpenMDW 1.1: A practical guide for companies adopting AI models
Memorised copyleft code in open-weight models and the obligations European Union law leaves undefined
Chardet: clean rooms, copyleft and the AI rewrite problem
Beyond the source code: the hidden licensing crisis in open AI
Beyond the source code: your infrastructure vendor changed its licence, and your legal team probably does not know
Beyond the source code: one GPL violation, five compounding legal exposures
Latest
Complete archiveOpenMDW 1.1: A practical guide for companies adopting AI models
OpenMDW 1.1 explained for companies: commercial use, outputs, third-party licences, termination rights and NVIDIA Nemotron.
Memorised copyleft code in open-weight models and the obligations European Union law leaves undefined
If open-weight AI models reproduce protected GPL code, EU copyright law may create copyleft obligations that companies cannot identify or perform.
Is the model the infringement? Memorisation, reproduction and the weights of AI models under European Union law
Two Munich courts have ruled that AI model weights can reproduce memorised works. A causal reading of EU copyright law, tested against the US and UK positions.
Chardet: clean rooms, copyleft and the AI rewrite problem
Chardet 7.0.0 was rewritten by AI and released without copyleft. Across the US, EU and UK, the rewrite likely escapes the LGPL: and that may be the real problem
If you are working through one of these, I would be glad to hear about it.
Get in touchStay in the conversation
New analysis sent when it is published.