Close Pandora's Box? To standardize the development of AI, countries have taken action

On April 11, the Cyberspace Administration of China (CAC) issued the Measures for the Administration of Generative Artificial Intelligence Services (Draft for Comments), which is the first regulatory document in the field of generative artificial intelligence (AI) in China. On the same day, the United States also reported that measures were being considered to regulate artificial intelligence. According to the Wall Street Journal reported on the 11th, the US Department of Commerce on Tuesday officially solicited opinions on artificial intelligence accountability measures, including whether potentially risky new artificial intelligence models should pass the certification process before release. In addition, the United Kingdom, Italy, Canada and other countries have also announced the tightening of artificial intelligence regulatory policies.

Infographic

Close Pandora's Box?

The Draft contains a total of 21 measures, emphasizing the premise and necessity of security assessment and algorithm filing. The Draft clearly states that generative AI as used in the Measures refers to technologies that generate text, images, sounds, videos, codes, etc. based on algorithms, models, and rules, and puts forward multiple requirements for generative AI services, including "the provision of generative AI products or services shall comply with the requirements of laws and regulations", "providers shall be responsible for the legality of pre-training data and optimization of training data sources of generative AI products", and "require users to provide real identity information".

A number of experts said in an interview with the Global Times on the 12th that the release of the Draft for Comments is mainly to clarify the general tone of generative artificial intelligence first standardizing and then developing, in order to improve the application environment of generative artificial intelligence and create conditions for the application of generative AI in China.

Recently, generative artificial intelligence technology has set off a new round of technological waves of productivity innovation, and major technology companies around the world are vying for layout. However, while the popularity of artificial intelligence is spreading, the conceptual chaos and security issues caused by it have also caused many concerns.

U.S. law professor Turley said in USA Today on April 3 that he received an email telling that someone found that he inexplicably appeared in the list of sexually harassing students given by the tool when testing generative artificial intelligence; Reuters reported on the 5th that Hood, the county governor of Hepburnshire, Australia, was mistakenly listed by generative artificial intelligence as the "perpetrator" of a foreign bribery scandal in a subsidiary of the Reserve Bank of Australia... Many say AI has opened a "Pandora's box" and called on governments to introduce regulatory measures.

In addition to China and the United States, the United Kingdom, Italy, Canada, France, Spain and other countries have also begun to act. According to the Financial Times, the UK government released its first AI white paper in March, outlining 5 principles of AI governance. In addition, at the end of March, the Italian government announced a ban on the use of an artificial intelligence chatbot and restricted the processing of Italian user information data by related companies, while opening an investigation into its privacy and security issues, Reuters reported. The Canadian Broadcasting Corporation reported on April 4 that Canada's federal privacy regulator announced that it has launched an investigation into OpenAI, a U.S. artificial intelligence research company, for allegedly "collecting, using and disclosing personal information without consent."

In contrast to many countries that have stepped up AI regulation, India has no plans to regulate the use and development of AI for the time being. According to Fox News, India's Ministry of Electronics and Information Technology issued a statement last week detailing India's position on AI development and safety, expressing confidence that AI will drive the growth of Indian business, and the government is taking all necessary measures in terms of policy and infrastructure to develop the country's AI industry.

Tech giants are also debating standards for AI development

Normally, the tightening of government regulation means that the space and power of large technology companies to develop freely are curbed, which will attract industry opposition, but this time, many technology companies have spoken out to call on the government to set regulations, and even within some companies, restrictions on the use of artificial intelligence have been implemented.

According to South Korea's "National Daily", a number of Korean companies, including Samsung and SK Hynix, are taking measures to restrict the use of relevant artificial intelligence tools to prevent employees from leaking secrets. LG Electronics also requires employees to conduct training on "artificial intelligence and data use." In Japan, Japanese companies including SoftBank, Hitachi, Fujitsu and Sumitomo Mitsui Banking Corporation have begun to restrict the use of interactive AI services in their business operations due to concerns about information leakage.

Some AI companies are also actively seeking to develop new regulations. According to the Wall Street Journal reported on the 11th, "We believe that powerful artificial intelligence systems should undergo rigorous security assessments", OpenAI said in a recent article: "Regulation is needed to ensure that safety assessments are adopted, and we are actively engaged with governments to explore the best form of regulation that may be adopted." ”

Fox News reported that on the 12th, Microsoft, Google, Apple, NVIDIA and other leading companies in the field of artificial intelligence development will discuss the development and use of AI technology standards and discuss how to continue to develop artificial intelligence in the most responsible manner.

"Finding a balance between development and security"

In fact, the "flame" of AI regulation has not been ignited recently. Back in April 2021, the European Commission presented the draft Regulation on the Artificial Intelligence Bill, which is seen as a milestone event for the EU in the field of artificial intelligence and the wider EU digital strategy. However, the progress of the proposal has not been as smooth as expected, and MEPs have not yet agreed on the basic principles of the proposal.

The National Institute of Standards and Technology also released version 1.0 of the AI Risk Management Framework (AI RMF) in January this year, which aims to guide organizations in developing and deploying AI systems to reduce security risks, avoid bias and other negative consequences, and improve the credibility of AI.

Shenyang, professor of the School of Journalism and Communication of Tsinghua University and director of the Metaverse Culture Laboratory of Tsinghua University, said in an interview with the Global Times on the 12th that all countries attach great importance to the regulatory issue of artificial intelligence, because artificial intelligence technology has greater uncertainty and uncontrollability. At present, it is necessary to guide the development of artificial intelligence industry to a safe and correct path to prevent major black swan events from occurring, causing irreparable losses.

For the future regulatory prospects of artificial intelligence, Pan Helin, co-director and researcher of the Digital Economy and Financial Innovation Research Center of Zhejiang University International Business School, told the Global Times that in terms of legislation, China is currently at the forefront, and the content of the Draft will continue to be improved in practice. He also said that in the future, supervision should find a balance between the development of generative AI and safe and orderly promotion, first, to give appropriate inclusion to AI-generated content; The second is to continue to refine the rules, which will greatly enhance the certainty of corporate behavior and be more conducive to the development of the generative AI industry.