JAKARTA - The content licensing company for music, images, videos, and other datasets used in artificial intelligence system training has formed the first trading group in the sector, announced on Wednesday, June 26.
The Dataset Provider Alliance (DPA) will advocate "an ethical data source" in AI system training, including the rights of people described in the dataset and intellectual property rights protection of content owners, according to a statement from the company.
Founder members include US music dataset company Rightsify; image licensing service, viceual; Japanese stock photo provider Pixta; and German-based data market Datarade.
The emergence of a generative AI technology that has been able to mimic human creativity in recent years has sparked protests from content creators and a series of copyright lawsuits against tech companies such as Google, Meta, and ChatGPT maker OpenAI, backed by Microsoft.
Developers have trained the model by giving them a large amount of content, many of which are taken from the internet for free without the consent of those who create the work or have the right to it. Tech companies claim that this use is legal, but also secretly pay for access to private content collections to meet certain data needs and avoid legal and regulatory risks.
The prospect that licensed data requests will increase if copyright owners win in their legal battle has prompted the emergence of a new industry of companies that pack content and sell access to it for use by AI systems. As a result, groups have been set up to set ethical standards for such trades, such as Fairly Trained, a non-profit organization founded this year that certified models that do not use copyrighted material without a license.
SEE ALSO:
The DPA targets content from the transaction, requiring its members not to sell text data obtained by exploring webs or audio featuring people's voices without their explicit consent. The main focus will drive legislation such as the NO FAKES Act, a US law introduced last year to create sanctions for the creation of voice digital replicas or unlicensed similarity, said Alex Bestall, CEO of Rightsify and its subsidiary, GCX, leading the group's founding.
"Advocacy will be a big part of this as everyone has taken their position on AI and copyright, but many of these battles have not been resolved and will take time to resolve them," Bestall said.
"The DPA will also encourage the transparency requirements of more training data, such as in the European Union AI Act and a similar US law introduced in April, namely the Generative AI Copyright Disclosure Act," he added.
"We plan to issue a white paper outlining our position in July," he said.
The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)
