Sony Reports Removal of Over 75,000 AI Deepfake Songs

Sony is currently dealing with the surge of AI-generated music that imitates some of its top artists, such as Harry Styles and Beyoncé. The company has reported the removal of over 75,000 tracks from online platforms that replicate the sounds of these artists. However, they believe this number represents only a small fraction of deepfake music available online.
The rise of deepfake music poses a significant threat to legitimate recording artists, leading to financial losses, as emphasized by Sony in its feedback to the UK government. The government is contemplating new copyright laws that would allow AI models to be trained on existing artists’ works. This development raises concerns among creators about the protection of their intellectual property.
Generative AI, which includes various tools like chatbots and music creation software, is still refining its capabilities. While text generation can often lead to inaccuracies and nonsensical outputs—such as making false statements—audio and image generation tend to require less precision. For instance, a dog’s appearance may vary, but a chatbot should not misstate basic math like “1+1 = blueberry.” Advocates argue that AI-generated content can help reduce production costs, yet they acknowledge that human creativity remains crucial for telling compelling stories and crafting narratives. There are fears that industries might produce a higher volume of low-quality entertainment, as studios focus on cutting costs to boost profit margins in streaming.
There is an ongoing conversation about whether listeners can discern between AI-generated songs and those created by human artists. The release of a song featuring synthetic voices mimicking Drake and The Weeknd highlighted concerns that listeners might not care about the origins of the music. This situation raises alarms about a future where genuine artists become less prominent, and streaming platforms are flooded with AI-generated tracks lacking originality.
In the UK, Prime Minister Sir Keir Starmer has expressed a desire for the nation to lead in AI innovation. He proposed that AI companies should be allowed to train their models using a variety of content, including music, without charge. However, this plan might require companies like Sony to actively opt out, which they see as an added burden in an already challenging environment.
Despite some artists opting to allow their likenesses for AI use, this is not the standard practice among musicians. In the UK, there have been weeks of protests surrounding the government’s new proposals, with artists worried that enforcing copyright laws would become increasingly complex under these rules.
A more pressing issue concerning AI technology is the creation of explicit deepfake content, particularly in the form of manipulated images where real individuals’ faces are placed onto explicit bodies, creating realistic yet damaging results. This problem has surfaced notably in high schools across the United States, illustrating a serious ethical dilemma. Meanwhile, deepfake audio is being misused in phishing scams, further highlighting the potential dangers of this technology.