news

Nancai Compliance Weekly (Issue 153): Nvidia is under antitrust investigation by the US Department of Justice; Internet ID and Internet Certificate are planned to be implemented, and some APPs have started pilot projects

2024-08-05

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

21st Century Business Herald reporter Xiao Xiao reports from Beijing

This week (July 29-August 4), the overall compliance dynamics were relatively flat, but events in the field of artificial intelligence are worthy of attention.

In terms of personal information protection: Internet number and Internet certificate are planned to be implemented, and the APP has started a pilot program.

In terms of artificial intelligence: AI was found to have "cheated" in a photography competition in Sichuan, and the organizing committee disqualified the works from winning awards.

Overseas: Nvidia is subject to an antitrust investigation by the US Department of Justice; Suno admitted that the company used copyrighted music works to train AI; Perplexity launched an advertising revenue sharing plan and has reached agreements with many companies; the US Intellectual Property Office released the "Copyright and Artificial Intelligence Report".

1. Personal Information Protection

1. Internet number and network certificate are planned to be implemented, and the APP has started the pilot

On July 26, the Ministry of Public Security, the Cyberspace Administration of China and others drafted the "National Internet Identity Authentication Public Service Management Measures (Draft for Comments)", which will issue "Internet IDs" and "Internet IDs" to the public. It is hoped that this will minimize the number of Internet platforms collecting and retaining citizens' personal information beyond their scope under the pretext of implementing the "real-name system". Users can authenticate their Internet IDs and Internet IDs to Internet platforms that require real-name authentication without having to enter information such as their names and ID numbers.

Although it is still in the draft stage, you can already apply for an online account and certificate on 67 apps and scenarios such as WeChat, Taobao, and Xiaohongshu. The "National Network Identity Authentication App (Pilot Version)" has also been launched in multiple app stores.

2. Artificial Intelligence

1. A photography competition in Sichuan was found to have AI "cheated", and the organizing committee disqualified the works from winning the awards

Many netizens reported that the third prize winner of the Sichuan Photographers Association's online competition "Searching for 'Green'" was suspected to be AI synthesis, and the reason for the doubt was that it was too neat and did not look like a real shot. The organizing committee did not directly respond to the AIGC controversy, but issued a statement on July 28 that after investigation and research, it decided to cancel the work's eligibility for selection and award.

3. Overseas Trends

1. Nvidia is under antitrust investigation by the U.S. Department of Justice

According to foreign media sources, the U.S. Department of Justice will launch an antitrust investigation into Nvidia, mainly because Nvidia's competitors complained that Nvidia may abuse its market dominance when selling artificial intelligence (AI) chips and force cloud computing providers to purchase multiple Nvidia products.

A Nvidia spokesperson wrote in a statement to the media on the 1st: "Our competition is based on decades of investment and innovation, strictly abiding by all laws, making Nvidia products publicly available to all cloud services and local clusters of every enterprise, and ensuring that customers can choose the best solution for them."

In addition, Nvidia's plan to acquire Israeli startup Run:ai, reportedly for $700 million, has also attracted scrutiny from the U.S. Department of Justice, meaning Nvidia is facing two independent antitrust reviews in the U.S. Last Thursday, Nvidia fell 6.67%.

2. Suno admits that the company used copyrighted music to train AI

On August 1, music AI company Suno issued a statement in response to the lawsuit filed by the three major record companies (Universal Music, Sony Music, and Warner Music), admitting that it had used the record content of these three major record companies to train AI, but the company believed that this was in line with the principle of fair use.

Suno argued in its submission that copying copyrighted works is only an intermediate step in training AI, and that making “intermediate” copies is protected by law. The company’s CEO wrote in a post: Using data from the “open internet” to train AI models is no different than “kids writing their own rock songs after listening to rock music.” Learning is not an infringement. It was not in the past, and it is not now.

3. Perplexity launches advertising revenue sharing plan and has reached agreements with many companies

In response to the "plagiarism" controversy that has continued to ferment in recent days, the AI ​​search application Perplexity launched a publisher plan last week. The main contents include revenue sharing and the introduction of advertisements through the related questions function. When the publisher's content is quoted, the publisher will receive a share of the revenue.

The first partners revealed include Time magazine, German news site Der Spiegel, Fortune, Entrepreneur, The Texas Tribune and WordPress.com. Partners can access Perplexity's API for free and create custom answer engines on their sites that reference only that publisher's content.

4. The U.S. Intellectual Property Office released the Copyright and Artificial Intelligence Report

On July 31, the U.S. Intellectual Property Office released the "Copyright and Artificial Intelligence Report". The opinions on the report have been publicly collected since August last year. More than 10,000 creators submitted their ideas, and technology giants such as OpenAI, Microsoft, Meta, and Google submitted dozens of pages of idea descriptions.

The report states that it received a large number of public comments last year, hoping to "protect artists whose styles are imitated by AI." However, the US Intellectual Property Office believes that the current legal remedies are sufficient to deal with the impact in this field. For example, if a user enters the prompt "in the style of..." and the work that AI finally spits out not only copies the artist's style, but also has specific iconic elements, the US Intellectual Property Office believes that the infringement claim should be supported. In addition, some AI developers have set restrictions in large models that cannot generate the style of living artists. The "Report" also points out that this is an appropriate protection method.