How to regulate AI? Experts suggest improving the legal guarantee and standard system

Author: Xu Cihao

Source: Financial Associated Press

‌Image source: Generated by Unbounded AI tools

The AI upsurge triggered by OpenAI is continuing, but while AIGC-related applications improve production efficiency, they are also used by criminals to generate false content and disrupt the normal operation of society.

Guosen Securities Research Report pointed out that domestic AI applications are gradually expanding, and AI-related applications will gradually increase. In view of the risk of AI abuse, data security, cryptography, and AI supervision are all important technological development directions.

A few days ago, Meiya Pico released a new product that can detect and recognize videos, images, and text generated by AI. In fact, relevant departments in our country have been exploring the legal supervision of AIGC products and applications. Relevant experts also suggested that only by establishing and improving the relevant legal system can we fundamentally promote the healthy and sustainable development of my country's AIGC industry.

Meiya Pico Releases AI Regulatory Products

On May 15, Meiya Pico released a record sheet of investor relations activities.

The record sheet shows that in terms of AI supervision products, the company has released the AI-3300 "Huiyan" video image authentication workstation, which can monitor and identify deeply synthesized and forged video images and AIGC-generated video images. The AIGC detection platform launched by the company supports the detection and recognition of AIGC-generated text.

At the same time, Meiya Pico said that follow-up companies will continue to conduct research on AI security and regulatory fields including anti-generative AI, and deploy related products in a timely manner according to business needs.

A reporter from Metaverse NEWS called Meiya Pico, and the person in charge of the company said that from the perspective of business model, the company’s AI security products mainly serve regulatory agencies such as public security and network information, aiming at the use of generative technology on the Internet to conduct illegal activities. Information dissemination behavior identification and identification; in addition, security and compliance evaluation services can be provided for manufacturers that provide generative services.

The above-mentioned connection person said frankly that domestic large-scale model manufacturers are still in the stage of release and market application. From the perspective of the B-side, cooperation in the business field is still under discussion.

"With the maturity of deep counterfeiting technology and AIGC technology, the technological threshold for counterfeiting related video, audio, and images has been greatly reduced, and the risk of AI technology abuse has further expanded, and supervision needs to be strengthened urgently." Angel investor and senior artificial intelligence expert Guo Tao Xiang Yuan A reporter from Cosmos News said that AI monitoring tools have relatively good results in monitoring and identifying pictures, videos, and videos made using deep forgery technology and AIGC technology, which can reduce the risk of technology abuse.

Dr. Li Sanping, a senior executive scientist of Dell Technologies China R&D Group, told the Yuanverse News reporter that from a technical point of view, there have been some attempts to generate pictures and videos with AI. For example, Google will mark every AI-generated image created by its tools.

"It is technically easy to identify and does not affect users in use. However, the content of text is still being explored." Li Sanping said.

AI supervision is always on the way

Guosen Securities believes that as domestic Internet companies such as Baidu, 360, Ali, Huawei, and Tencent release or plan to release large-scale models, AI-related applications will gradually increase. In view of the risk of AI abuse, data security, cryptography, and AI supervision are all important technological development directions.

In fact, my country has always attached great importance to the supervision of AI.

On January 10, 2023, the "Regulations on the Administration of Deep Synthesis of Internet Information Services" (hereinafter referred to as the "Regulations") came into effect. The "Regulations" clarify that services that provide functions such as intelligent dialogue, synthetic human voice, face generation, and immersive virtual scenes that generate or significantly change information content should be marked prominently to avoid public confusion or misidentification. No organization or individual may use technical means to delete, tamper with, or conceal relevant marks.

On April 11, the Cyberspace Administration of China issued the "Measures for the Administration of Generative Artificial Intelligence Services (Draft for Comment)" (referred to as the "Measures"), focusing on generative AI for the first time, and further clarifying the legal boundaries and responsible entities. The "Measures" propose that organizations and individuals that use generative artificial intelligence products to provide services such as chat and text, image, and sound generation, including organizations and individuals that support others to generate text, image, and sound by themselves by providing programmable interfaces, etc., should Assume the responsibility of the producer of the content generated by this product. Providers should be responsible for the legitimacy of the pre-training data and optimized training data sources of generative artificial intelligence products.

"In terms of AI industry supervision, there are outstanding problems such as incomplete relevant laws, regulations and standard systems, lack of a top-down supervision system coordinated by multiple departments, and backward supervision technology." Guo Tao pointed out that relevant government departments need to establish and improve the AIGC R&D and Apply relevant laws, regulations and standard systems, establish or improve the ethical review system and supervision system, strengthen the supervision of technology companies, application scenarios, and users, build a multi-party collaborative governance platform, and promote the healthy and sustainable development of the AIGC industry.