The United States and China and the race for the rules of artificial intelligence, for opposite reasons

The United States and China and the race for the rules of artificial intelligence, for opposite reasons

yesterday I wasThe Chinese government has presented a document containing rules to regulate the use of generative AI, or those capable of producing text and images, such as ChatGPT and MidJourney. Last April, the Cyberspace Administration of China had already submitted a draft regulation: compared to the initial document, the new text provides a slight easing of requirements for companies in the sector and an affirmation of Beijing’s willingness to “promote the use of AI innovation in all sectors”.

To understand the content of the document on artificial intelligence, it is necessary to recall another initiative taken by the Chinese government last January, when it published the “Deep Synthesis Management Regulations for Internet Information Services”. In this case, there was a set of rules to limit the production of “synthetic media,” such as deepfakes, for example, a term used to refer to digitally produced or manipulated audio and video content, where features or voices of people are already inserted into private. Born into pornography, deepfakes quickly caught on becoming a fairly simple and cheap practice, but also effective at generating confusion and misinformation. In this regard, Chinese law provides restrictions for similar software producers, who will have to report the “synthetic” nature of this type of content.

We must therefore start from here to understand the document published yesterday, according to which generative AI programs must comply with administrative laws and regulations but also “respect morality and social ethics” and above all “adhere to the fundamental values ​​of socialism”. Finally, AI systems are prohibited from “incitement to subvert state power.” This means that no matter what the user prompts (the sentence with which you describe to the AI ​​what kind of text you want to produce), the end result of these services has to fall within the narrow confines of Chinese censorship.And the. To achieve this, the CAC has asked companies in the sector to “adopt effective measures” in the training and design phase of these algorithms, a point considered so important that it precedes those related to the respect of intellectual rights and copyrights.

See also  Meloni is in Washington, her eyes fixed on the Chinese note. Sisci writes

The Chinese regulation was published in the same hours that the Federal Trade Commission, a US agency that deals with consumer protection and combating anti-competitive business practices, announced that it had opened an investigation against OpenAI, the maker of ChatGPT.. Allegations against the company are varied, including failure to comply with consumer protection rules and liability for a personal data leak that occurred months ago. In the same week, Elon Musk, the co-founder of OpenAI and now a public enemy of his former partner Sam Altman, founded xAI, an industrial company that claims to want to “understand the universe” with artificial intelligence. While Western countries are quick to set rules to protect users from misinformation, China appears to be doing exactly the opposite: to ensure that propaganda works.

Chinese initiatives in this area have been sparking debate for some time: Democratic Senator Mark Warner declared last month that “China has a huge advantage in terms of regulating AI,” recalling how the comparison with the Asian country fared. This technology. A few days ago, Republican Senator Ted Cruz took a similar position and identified the possibility of Beijing surpassing the United States in the field of artificial intelligence and regulation as “extremely dangerous.”

As pointed out by the Carnegie Endowment for International Peace, a think tank, Beijing’s sharp and rapid moves in the field of artificial intelligence Their goal is to control information And in particular, algorithms, which are already the subject of specific regulation similar to that of deepfakes and generative artificial intelligence, dating back to 2021. A new national registry is dedicated to algorithms of this kind in which Chinese companies must register their creations, which will allow the government to collect information about operation and design.

See also  Perseids: The best show in recent years to predict

Leave a Reply

Your email address will not be published. Required fields are marked *