My next question is that given that these conversational assistants, as you stated, are based on large language models, LLM, how do you interpret their risk-based classification under the EU AI Act?

While not necessarily labeled as high-risk by default, generative AI systems, such as large language models, are recognized under the AI Act as a distinct classification scale.

The Act requires the respect of certain transparency obligations, for example, disclosing that content was AI-generated to the user, And also of certain compliance measures when these systems are deployed in high-risk contexts, including large language models being used in legal decision-making.

So the AI Act also introduces a new tier for general purpose AI models.

These are large-scale AI models trained on vast amounts of data that can be adapted to a wide range of downstream tasks, such as language generation, translation, or image recognition.

Popular examples are OpenAI's ChatGPT and Microsoft's Copilot.

The AI Act mandates rigorous documentation, risk assessments, and data governance protocols to ensure safety, reliability, and accountability.

According to your article titled Generative AI systems in legal practice, offering quality legal services while upholding legal ethics, does this raise concerns about lawyers maintaining the necessary level of diligence?