New regulations from the European Union (EU) came into effect on August 2, strengthening obligations for generative artificial intelligence (AI) models like ChatGPT and Gemini. The legislation focuses on ensuring that AI respects copyrighted content and does not provide discriminatory information or information that infringes on fundamental rights.

Key Provisions of the AI Act 

This regulation is part of the AI Act, which the EU began implementing in stages last year, with its application set to expand until 2027. This law is the world's first comprehensive AI regulation, allowing or prohibiting the use of AI technology based on its level of risk to people.

The European Commission recently published guidelines that generative AI systems must follow to comply with the law. These guidelines take the form of a voluntary code of conduct and include the following key points:

Copyright Protection: Companies are advised not to use copyrighted data to train AI models.
Transparency and Accountability: AI models must assess the risks they may pose to users and analyze the potential to cause serious harm, such as loss of life or chemical, biological, or nuclear accidents.
Anti-Discrimination: AI must not provide content that infringes on fundamental human rights or promotes discrimination.
 

Big Tech's Response and the EU's Stance 

Interestingly, some major tech companies have had mixed reactions to this voluntary code of conduct. Meta has stated it will not sign the code, and Elon Musk's xAI plans to sign only parts of it.

Thomas Regnier, a spokesperson for the European Commission on technological sovereignty, commented that "signing even a part of the voluntary code is a welcome step." However, he emphasized that "companies will have to prove compliance with the law through other means," highlighting that ultimately, adherence to the AI Act is essential. This move, along with other EU laws targeting large tech companies (the Digital Markets Act and Digital Services Act), is interpreted as an effort to curb the power of these tech giants.

EU Ushers in a New Era of AI Regulation: Proactive Measures to Protect Copyright and Fundamental Rights 

On August 2, 2025, the European Union (EU) officially activated a new legal framework to regulate the massive wave of artificial intelligence (AI) technology. The regulations coming into force demonstrate the EU's firm commitment to balancing technological progress with social responsibility by imposing strict obligations on generative AI models. The core of this legislation is to prevent AI models from using copyrighted content without authorization and from producing information that infringes on individual fundamental rights or promotes discrimination.

This measure is based on the EU's AI Act, which has been implemented in stages since last year. The law is considered the world's first AI regulation that permits or prohibits the use of AI systems based on the level of risk they could pose. The European Commission recently released a "voluntary code of conduct" detailing the specifics of this law, providing concrete standards that AI companies must follow.

According to this code, AI companies should not use copyrighted data during the model training process. This is a clear signal to protect the rights of creators and copyright holders. Additionally, AI models must thoroughly assess the potential risks they may pose to users. This includes analyzing the possibility of human rights violations as well as causing serious real-world dangers like chemical, biological, or nuclear accidents.

However, not all big tech companies are complying with this regulation. Meta, the parent company of Facebook and Instagram, has stated it will not sign the voluntary code, and Elon Musk's xAI has expressed its intention to sign only certain provisions. This suggests that the EU's regulatory movement may clash with the interests of U.S. tech companies. The U.S. has already expressed concerns about the EU's Digital Markets Act (DMA) and Digital Services Act (DSA), arguing that they could hinder technological innovation.

In response, the European Commission is taking a flexible yet firm stance. Thomas Regnier, the Commission's spokesperson for technological sovereignty, stated that "signing even parts of the voluntary code is a welcome development," but stressed that "in the end, AI companies will have to prove their compliance with the law through other means." He made it clear that while the code is voluntary, adhering to the AI Act is ultimately an obligation for all companies.

All new AI models entering the market must immediately comply with these regulations. If they fail to do so, the European Commission can initiate a formal investigation into the company starting August 2, 2026. However, the application of the regulations to existing AI models that are already commercialized has been deferred for one year. Despite this, over 40 European artists, publishers, and copyright organizations have expressed dissatisfaction with the code, claiming it is "insufficient to protect intellectual property rights." They argue for stronger, more binding measures to protect the rights of creators in the age of AI.

The EU's AI regulation will be a challenge and a new benchmark for tech companies. It introduces a new paradigm that AI technology should not just be a tool for generating economic profit but should also respect social values and fundamental human rights. In the future, AI technology will be evaluated not only on how powerful it is, but also on a new scale of how ethical and safe it is.