Spain’s government has taken a major step toward regulating artificial intelligence by approving a bill that imposes heavy fines on companies failing to label AI-generated content properly.

The legislation, introduced on Tuesday, aims to combat the spread of deceptive AI-created media, commonly referred to as “deepfakes.”

The proposed law aligns with the European Union’s landmark AI Act, which places strict transparency obligations on high-risk AI systems, Digital Transformation Minister Óscar López announced. “AI is a very powerful tool that can be used to improve our lives… or to spread misinformation and attack democracy,” López stated during a press briefing.

Spain is among the first EU nations to implement the bloc’s AI rules, which are considered more comprehensive than the largely voluntary regulatory approach taken in the United States.

The bill, which requires approval from the lower house of parliament, categorizes failure to label AI-generated content as a “serious offence.” Violators could face fines of up to €35 million ($38.2 million) or 7% of their global annual revenue.

Ensuring AI does not pose a societal threat has been a priority for regulators worldwide, especially since the emergence of ChatGPT in late 2022. The chatbot, developed by OpenAI, demonstrated AI’s ability to engage in human-like conversation and perform various tasks, raising concerns about potential misuse.

The bill also bans certain AI applications, including subliminal messaging that could manipulate vulnerable groups. Examples cited include chatbots that encourage gambling among addicts and AI-powered toys that push children toward dangerous behaviors.

Additionally, the legislation restricts organizations from using AI to classify individuals based on biometric data, such as facial recognition, to assess their eligibility for benefits or predict criminal behavior. However, authorities would still be permitted to use real-time biometric surveillance in public spaces for national security purposes.

Oversight of the new AI regulations will be handled by the recently established AI supervisory agency, AESIA. However, specific areas such as data privacy, crime, elections, credit ratings, insurance, and capital markets will remain under the jurisdiction of their respective regulatory bodies.

Spain’s proactive approach underscores the growing need for robust AI governance as technology continues to evolve, ensuring its benefits do not come at the cost of public trust and security.