German creatives want EU to fix ChatGPT copyright issue

ChatGPT went on a triumphant welcome tour in Europe. Follow

to complain about Italian regulators and

European Parliament, it is the turn of the German trade unions to express their concerns about potential copyright infringement.

No fewer than 42 trade organizations represent the country More than 140,000 authors and performers have signed a letter urging the European Union to enforce strict rules on AI’s use of copyrighted material.

first by Reported Reuters, this letter highlights people’s concerns about the materials used to train large languages The Growing Concern Model (LLM) of Copyright and Privacy Issues, Statement,

“Unauthorized use of protected training material, its opaque handling, and foreseeable alternative sources of output from generated AI raise fundamental questions of accountability, liability, and remuneration that need to be addressed before irreversible damage occurs “

Signatories include Germany’s main trade unions Verdi and DGB, among others Photographer, designer, journalist and illustrator. The author of the letter further added that

“Generative AI requires At the center of any meaningful AI market regulation.”

ChatGPT is not The sole target of copyright fights. In January, visual media company Getty Images filed a copyright claim against Stability AI. According to the lawsuit, the image maker developer allegedly copied more than 12 million photos, captions and metadata without permission.

LLM training with diminishing returns

The arrival of OpenAI’s ChatGPT It sparked a wave of concern. So far, these have run the gamut from aggressive development resulting from a commercially motivated AI “arms race” to privacy, data protection, and copyright concerns. The latest model, GPT-4, was trained using over a trillion words.

At the same time, one of the instigators of the controversy, the company’s CEO Sam Altman, on Zhou said the amplification machine behind ChatGPT Learning strategies have to go with the flow. In fact, OpenAI predicts diminishing returns from scaling up the model. The company spent about $100 million to train its latest model, GPT-4, using more than 1 trillion words.

Meanwhile, the European Union’s Artificial Intelligence Act Tension close to landing. While it will likely set global regulatory standards, the question is how adaptable it will be as developers find other innovative ways to make algorithms more efficient.