The new US government wants to change how artificial intelligence is handled in the United States. One of the biggest players in this field is now resisting and saying: If their AI is no longer allowed to read protected content, we would lose a race against China.
What is OpenAI?
- OpenAI was founded in 2015 and researches artificial intelligence.
- With ChatGPT, OpenAI has developed a generative AI that can generate texts and answer questions.
- The most famous face of the company is CEO Sam Altman (featured image), who only made headlines in January when he said AI could threaten our society.
Why does OpenAI need protected material? Artificial intelligence needs to be “trained”. This means that the machines take, for example – mostly texts created by humans – and learn from them how a text actually works.
The more material available, the better the AI becomes at what it does, whether it be images, videos, or texts. However, the AI does not care whether a work is protected or not. It learns and can, in some cases, reproduce an entire book without the author ever having been paid for it.
Already in 2023, there was a huge outcry among renowned authors. The father of Game of Thrones, George R. R. Martin has sued ChatGPT with 16 colleagues.
Without protected material, the race is effectively over
The current US government under President Trump plans a reform on how to deal with artificial intelligence (via whitehouse.gov). Suggestions and comments on how the USA can become a global leader should be submitted by March 15.
OpenAI has now put forth numerous ideas and shared them on the official blog. There it says:
The government can simultaneously secure Americans’ freedom to learn from AI and prevent handing leadership to the People’s Republic of China by maintaining the ability for American AI models to learn from protected material.
If it is restricted that artificial intelligences are allowed to read protected material, the race to develop AI is “effectively over” – because these restrictions do not exist in China. Besides that, as stated in the more detailed paper on this, smaller companies with limited budgets need a wide range of content to train their AI.
How much truth there is to these claims is difficult to say without solid knowledge of artificial intelligence. However, China has recently proven that the country is indeed competitive and even at a small price: China has just dealt a heavy defeat to the USA: While ChatGPT costs 700,000 euros per day, DeepSeek is available for 83,500 euros.
