Google recently announced a new partnership with Universal Music Group (UMG) to develop an “AI framework” for generating music content. This comes on the heels of the viral success of “Heart on My Sleeve,” an AI-generated song featuring the simulated voices of Drake and The Weeknd, two UMG artists. UMG was unhappy about this unauthorized AI imitation of their artists and demanded the song’s removal from platforms like YouTube.
The “Heart on My Sleeve” situation presented a tricky predicament for Google and YouTube. On one hand, YouTube risks losing its licensing deals with major music labels like UMG if it does not address AI-generated content that imitates artists without permission.
On the other hand, Google wants to continue training its AI systems on large amounts of web content, arguing this is permissible under fair use.
Critics worry the UMG partnership signals YouTube will expand its copyright enforcement technologies like Content ID to give labels more control over AI-generated content, even if it does not clearly violate existing laws.
This could allow UMG to claim new rights over artificial voices mimicking their artists without legal precedent. Smaller creators on YouTube would have little recourse against erroneous copyright claims.
At the same time, Google denies publishers the same leverage over how it uses their content to train AI systems. While using copyrighted songs to mimic artists raises legal questions, Google insists scraping news websites to train AI models is fair use. This puts publishers in a difficult position, as blocking Google’s access to their sites cuts off a major traffic source.
The growing battles over AI and copyright highlight the immense power held by tech giants like Google. With one hand, Google placates big media companies like UMG with lucrative deals over AI content.
With the other, it scrapes publisher websites to feed its AI, rejecting publishers’ requests for similar compensation.
As AI technology progresses, pressure mounts on legislators to update intellectual property laws for the digital age. Until then, smaller internet creators and publishers will likely find themselves squeezed between the interests of big tech and big media. The coming copyright disputes will shape the future of creativity and culture online.
While YouTube tries to appease the music industry, Google continues leveraging its dominance over web traffic to scrape data for its AI systems.
Publishers have little choice but to comply, even as Google rolls out features like Search Generative Experience (SGE) that could directly compete with them. SGE uses AI to answer search queries, often with transactional intent – like product recommendations.
Google maintains it can use any publicly available web content to train its AI models, as allowed under fair use. But publishers argue scraping their expensive, human-generated content to fuel AI that then replaces them in search results is hardly fair.
Unlike music labels, most publishers lack bargaining power over Google. Their only recourse is trying to legally challenge scraping for AI training as copyright infringement.
Some publishers like The New York Times have updated their terms barring the use of their content for AI training.
But even the Times stops short of technically blocking Google’s access in its robots.txt file. For now, publishers’ legal threats lack teeth compared to Google’s grip on search traffic. Content blocking means virtual irrelevance.
The brewing copyright disputes around AI threaten to reshape the social internet. Previously, “everything is a remix” was a cultural ethos. Now, creators seem to be moving toward more of a “fuck you, pay me” attitude.
The law lags behind technology, leaving room for confusion. Can our brains “train” on copyrighted material by reading without infringing? Publishing schemes like Google’s special UMG deal sidestep this debate.
In time, AI-generated content pollution could push Google itself toward licensing deals for high-quality training data. Its search results would essentially become another walled garden like YouTube. The open web Google once disrupted would be choked out.
The coming legal battles around AI copyright will stretch for years absent legislative action. Until then, power dynamics favor Big Tech over creators and publishers. Pushing for a more balanced AI future requires updating both the law and cultural attitudes on intellectual property.