Artificial intelligence has been reshaping computing for decades, but the rise of generative solutions has multiplied its legal challenges. Copyright, data protection, image rights, discriminatory bias and trade secrets are just some of the issues raised by these new models. The AI Act has begun to provide a framework, although many questions, including the responsibilities of the different actors involved and the exploitation rights over model outputs, remain open.
At Audens, we help ensure that AI-based projects (from high-risk systems to generative AI or general-purpose models) comply with the AI Act and all applicable regulations. At the same time, we anticipate legal risks in still-uncertain areas, integrating compliance from the design stage, even before selecting the model or beginning the training phase.
AI Act compliance
We advise on identifying and implementing the obligations set out in the AI Act, including documentation, transparency requirements and the design of appropriate oversight processes.
Contracts
We draft agreements to regulate AI projects from the earliest consultancy stages, avoiding ambiguities that may lead to disputes during development, training or subsequent exploitation.
Training and fine-tuning
We advise on the selection and use of data for training and fine-tuning models and systems, minimising risks of bias and of infringing intellectual or industrial property rights.
Impact assessments
We assess how high-risk AI systems may affect individuals' fundamental rights, identify associated threats and recommend measures to mitigate them.
Unlawful uses
We take action against AI uses that infringe rights (such as impersonation, deepfakes or the unauthorised exploitation of content) seeking their removal and, where appropriate, compensation.
Legal and ethical opinions
We prepare legal and ethical opinions on different AI uses, particularly to meet the requirements established in European grants and funding programmes.
Key experience
We assisted a multinational in the design and deployment of an internal AI system intended to identify the most suitable professionals for each project. Our work included data-protection and AI-risk assessments and the definition of measures to enhance transparency, avoid bias and ensure effective human oversight.
We advised a company specialised in dermatology-focused AI. Our support included anonymising clinical data for model training, negotiating agreements with pharmaceutical companies and clinics to enrich the dataset, and providing compliance training to their teams in both data-protection and AI-regulation requirements.
We served as ethical advisor to an AI-based patient-monitoring project funded under the EU's EIC Accelerator programme. Our opinions assessed risks related to data protection, patient autonomy and human dignity, proposing mitigation measures that enabled compliance with the standards required by European authorities.
We deliver professional training on the AI Act for Aranzadi La Ley, with several editions already completed and hundreds of participants. These sessions allow us to share our hands-on experience with the new regulatory framework.
