The European Union has implemented the Artificial Intelligence Act (AI Act), which enters a new phase of application as of August 2nd. This regulation introduces significant obligations regarding transparency, protection of citizen rights, and supervision of AI systems, impacting thousands of companies, professionals, and public bodies, including those in Granada.
Penalties for non-compliance can be substantial, reaching up to 35 million euros or 7% of global annual turnover for serious violations like prohibited practices. Fines for breaching transparency obligations can amount to 15 million euros or 3% of turnover. The European AI Office will be responsible for monitoring, investigation, and imposing sanctions.
The main novelty of the law is the increased obligation for transparency. Users must be informed when interacting with an AI chatbot, and certain AI-generated content (videos, images, audio, photographs) must be identified if it could be mistaken for real material. This particularly affects deepfakes, aiming to combat disinformation and build trust in these technologies. However, exceptions exist for artistic, satirical, humorous, or fictional works, provided it is evident to the user that they do not represent real events.
The regulation is not solely aimed at tech giants like OpenAI or Google, but at any company, administration, or professional incorporating AI systems into their daily activities, regardless of size. In Granada, this includes communication agencies, law firms, clinics, businesses, and public administrations using AI for various tasks, from campaign design to candidate selection.
Organizations will need to review their practices, understand the AI systems they employ, inform users about AI interactions, identify AI-generated content that could be misleading, and establish internal procedures for employees' use of these tools. Training on responsible use will also be promoted, and system traceability will be ensured.
The European Artificial Intelligence Act applies to AI systems marketed or used within the European Union, affecting popular tools such as ChatGPT, Gemini, Claude, Microsoft Copilot, Midjourney, or DALL·E, among others. Requirements will vary based on the risk level associated with each system and its application.
The legislation prohibits sensitive AI uses, including manipulating human behavior, certain real-time biometric recognition systems, and generating fake intimate imagery without consent. It aims to protect minors and vulnerable groups. The EU positions itself as the first major region with a specific legal framework for AI, seeking to balance innovation with fundamental rights.
