Can deepfake be art? AI's role in the creative industry

While the term "deepfake" often carries negative connotations due to its association with the unauthorized use of individuals' likenesses, the technology behind it holds significant potential as a legal and valuable tool within the creative sector. This article examines how machine learning techniques can support film, music, and other cultural productions, while also addressing the regulatory landscape, particularly the EU's AI Act, concerning the disclosure of AI-generated content.

“

Deepfake means an image, audio, or video content generated or manipulated by AI, which resembles existing people, objects, places, entities, or events, and in reception may be wrongly considered authentic or true.

The pervasive advancement of technology is increasingly evident in the realms of culture and art. AI-powered tools offer creators a wide array of possibilities. Deepfake content, which involves AI-generated or manipulated images, audio, or video resembling real people, objects, places, or events, is no longer confined to personal likenesses but is finding applications in various creative fields.

The EU's AI Act introduces crucial transparency requirements. Starting August 2, 2026, entities employing AI systems that generate or manipulate deepfake content must disclose that the output is AI-generated. This intervention must be easily detectable, and users interacting with such systems must be informed. The European Commission is further refining these obligations through a second version of the Code of Conduct on Artificial Intelligence-generated Content, emphasizing that disclosure should accompany content wherever feasible. Multimodal content containing deepfakes must be consistently labeled with an icon or tag, clearly visible to users without requiring further action.

“

According to the AI ACT, every entity using an AI system that generates images, audio, or video content constituting deepfake content or manipulates such content is obliged, among other things, to disclose that the content is the result of artificial intelligence.

This regulatory push raises questions about the balance between transparency and creative freedom. The article touches upon the potential for these restrictions to infringe upon fundamental human rights, specifically the freedom of creation, suggesting a complex interplay between technological advancement, ethical considerations, and artistic expression.

“

All multimodal content containing deepfake should be consistently disclosed using an icon or label, ensuring that this disclosure is clearly visible to the natural person without requiring further action from them.