Core Viewpoint - Deezer is taking significant steps to combat streaming fraud by labeling AI-generated music, as approximately 18% of daily uploaded music is now fully AI-generated, with 70% of its plays being fraudulent [1][2]. Group 1: AI-Generated Music and Streaming Fraud - Deezer will start labeling albums that contain AI-generated tracks to address streaming fraud [1]. - Over 20,000 tracks uploaded daily are fully AI-generated, with 70% of their play counts being fake [1]. - Currently, AI-generated songs account for only 0.5% of total traffic on Deezer, but this trend is rapidly increasing [1]. Group 2: Transparency and Artist Rights - Deezer emphasizes the importance of responsible and transparent practices to build trust with users and the music industry [2]. - The company has applied for two patents for its AI detection technology, which focuses on identifying unique features to differentiate synthetic content from real content [2]. - Deezer is committed to protecting the rights of artists and songwriters amid ongoing copyright law challenges related to AI model training [2].

Core Viewpoint - The global recorded music market reached a milestone in 2024, with total revenue soaring to $29.6 billion, driven significantly by streaming, which surpassed $20 billion for the first time. However, the rise of AI technology has led to an increase in streaming fraud, making it a target for criminals [1]. Group 1: AI-Generated Music Fraud Case - A notable case involved a 52-year-old man from North Carolina, Michael Smith, who was charged with using AI to create fake songs and fraudulently distribute them on platforms like Amazon Music, Apple Music, Spotify, and YouTube Music to illegally obtain royalties totaling over $10 million [3][4]. - Smith's fraudulent activities began in 2017, where he created thousands of automated accounts to continuously play AI-generated songs, resulting in billions of plays and substantial royalty income [4][5]. - The indictment revealed that Smith purchased hundreds of thousands of AI-generated songs and manipulated streaming data to create the illusion of global listening [5][6]. Group 2: Impact of AI on Streaming Fraud - The use of AI in music generation has allowed fraudsters to upload millions of fake songs to streaming platforms, with individual songs receiving only a few thousand plays, making it harder for platforms to detect fraud [8][9]. - Deezer reported that 18% of the content uploaded to its platform is generated by AI, highlighting the scale of the issue [9]. - The International Federation of the Phonographic Industry (IFPI) noted that AI is a significant driver of streaming fraud, enabling criminals to evade regulation while achieving profitability [10][11]. Group 3: Legal and Industry Responses - Legal actions have been taken against individuals involved in music streaming fraud, with a notable case in Denmark resulting in a prison sentence for a man who earned approximately 2 million Danish Kroner (about $229,676) through similar fraudulent activities [12][13]. - Despite efforts by platforms like Deezer to combat AI-generated music fraud, the effectiveness of these measures remains questionable, as the problem continues to escalate [15][16]. - There is currently a lack of legal frameworks governing the distribution of AI-generated music, complicating the enforcement of copyright protections [17].

Core Insights - The rise of AI-generated music is significantly impacting the music industry, with platforms like Deezer reporting that 18% of newly uploaded content is entirely AI-generated, up from 10% in January 2023 [1][5][11] - The International Confederation of Societies of Authors and Composers (CISAC) estimates that by 2028, AI could erode up to 24% of music creators' income [1][11] - Various music platforms are developing tools to detect AI-generated content, with Deezer and YouTube leading the way in implementing detection systems [4][7][14] Industry Response - Deezer has introduced a tool to detect AI-generated music, claiming it can identify 100% AI-generated tracks and is working on a system that does not require specific datasets [7][12] - YouTube has added a "synthetic voice recognition" feature to its Content ID system to automatically detect AI-generated content [7][14] - The Korean Music Copyright Association (KMCA) has implemented a strict "0% AI" rule for song registration, while the U.S. Copyright Office allows registration of AI-assisted works with significant human involvement [11][12] Challenges in Detection - Despite advancements in detection tools, identifying AI-generated music remains a challenge, as human listeners struggle to differentiate between AI and human-created music [4][19] - A study indicated that the average accuracy for identifying AI-generated music was only 46%, with even professionals only achieving a 50% accuracy rate [19][21] - The prevalence of AI-generated music raises concerns about the potential saturation of the market with low-quality, soulless music, as expressed by artists like Grimes [7][11] Legal and Ethical Considerations - The music industry is grappling with legal issues surrounding the use of copyrighted material to train AI models, with ongoing lawsuits against AI music startups like Udio and Suno [16][21] - There is a philosophical debate about the nature of creativity and originality in the age of AI, questioning whether AI-generated music can evoke genuine emotions and what constitutes true creativity [21] - Prominent artists, including Paul McCartney, have voiced concerns about AI's potential to undermine musicians' rights and the integrity of their work [12][16]