Silent Albums And Pirated Books: The AI Copyright Showdown

Two major stories emerged this month highlighting tension climbing between the development of AI and creative rights – a conflict that could reshape the way content is created, detained and monetized for years to come .

The sound of silence

Yesterday, saw a significant development in the current tensions between the creators and the companies of AI as more than 1,000 musicians – including Kate Bush, Annie Lennox, Yusuf / Cat Stevens and Damon Albarn – released a completely silent album entitled “Do we want?”

The project, available to broadcast on Spotify, offers records of empty studios and performance spaces and represent what artists fear becoming a reality if British law changes on British copyright were pushed. All the benefits of the silent album will be given to musicians to help the charity.

The silent demonstration coincides with the end of a consultation on the proposals of the British government which would create a copyright exemption for the training of AI models. This exemption, if it was approved, would allow technological companies to use material protected by copyright without license, while obliging creators to be made actively to protect their work.

IA’s creative data hunger

These developments highlight a fundamental tension in the advancement of AI – advanced models require large amounts of training data, with creative work representing some of the most precious and sought -after inputs. For technological companies to develop to develop increasingly sophisticated AI systems, access to this content has become an increasingly important strategy – that which stimulates an increasing range of legal battles and political debates.

Recent court documents filed in January, for example, revealed that Meta allegedly torrential More than 80 teraoctes of pirated books of “Shadow libraries” like Libgen and Z-Library, to form its models of Llama with large language. The internal communications revealed also suggest that the meta-PDG Mark Zuckerberg has personally approved using Libgen equipment, despite the warnings of the company management team of the company it was ” Data set that we know how to be hacked “.

A message from a Meta employee, quoted in legal files by authors such as Ta-Nehisi Coates and actress Sarah Silverman, noted that “media coverage suggesting that we have used a set of data that we know to be hacked , as Libgen, can undermine our negotiation position with regulators.

Elsewhere, in October 2024, the New York Times sent a Notice to stop ceasing to perplexityaccusing him of using the content of the publication without authorization. This followed the previous trial of the New York Times against Openai and Microsoft for having allegedly stolen its items to form the Chatppt. The confrontation represents what Marc McCollum, director of innovation at Raptive, calls “A decisive struggle for the soul of the Internet.”

Economic issues

The economic implications of these conflicts are substantial. In the United Kingdom, music alone contributed 7.6 billion sterling pounds to the economy in 2023With exports reaching 4.6 billion pounds sterling. On a larger scale, creative industries represent the world of billions of economic activities and millions of jobs the world.

At the same time, the development of AI has become central to technological strategy, companies like Meta, Google, Openai and others invest billions in the progress of their models. The perceived competitive need for these investments could create enormous pressure to secure training data by all the available means.

This shock of economic interests lies at the heart of the protest of the silent album and the legal challenges to companies like Meta. Like Ed Newton-Rex, the British composer and former Director of AI who organized the silent album, sums up: “The government’s proposal would freely put the work of the life of the country to companies, allowing these companies to ‘exploit the work of musicians to exceed them. “

Meanwhile, the British Government Department for Sciences, Innovation and Technology argues that “the current copyright regime and AI prevents creative industries, the media and AI from being full potential”. The ministry insists that it seeks a balance that allows the two sectors to prosper, adding that no decision has been made “and” no movement will be made until we are absolutely convinced that we have A practical plan that provides each of our objectives “.

Defense of practices, defining the limits

Companies in the center of these controversies have developed various defenses for their data acquisition strategies. Meta’s legal argument depends on the characterization of the torrent as “a widely used protocol to download large files” from “a well -known online benchmark which was accessible to the public”, according to court documents. This approach tries to normalize what the authors describe as “one of the largest data hacking campaigns in history”.

Similar debates take place in the AI ​​landscape while companies are trying to define the limits of the use of legal and ethical data. Fundamental questions remain consistent in all industries – to whom belongs creative content? When is the use of creative content for AI training allowed? And what remuneration models are appropriate?

As for public opinion, it seems to support stronger protections for creators. A investigation By the News Media Alliance found solid public support to protect the media and content creators against the unfair use of their content by AI systems, with a majority of public support regulations if it affects livelihoods.

AI and creativity: the path to follow

The protest of the silent album and Meta’s legal problems are only two examples which represent different facets of the same fundamental tension between the advancement of AI and creative rights. “You are not promoting growth in a garden by allowing all pests to feast on fruit and flowers”, as Stephen Fry said in his criticism of British proposals “, and you do not favor the growth of an economy by allowing all ISs to feast on the fruits of our creators.

The resolution of these conflicts will probably imply a combination of precedents legal, political frameworks, technical solutions and new commercial models. As for creators who want their work to be paid enough for their efforts, their hope will probably be for an industry that evolves towards a more structured approach to the use of content for AI training. Although access without restriction to creative works can become less common, however, exactly how these new frames will take shape, remains uncertain.

While technological companies and creation professionals sail in this complex AI landscape, the decisions made in the following years could have profound implications – potentially to reshape the way in which we define the property and the creation of value to A time when machines can more and more imitate human creativity.