No menu items!

Biden signs ‘AI risk prevention’ executive order… Safety test results must be reported

Share This Post

- Advertisement -
U.S. President Joe Biden is signing an executive order to regulate artificial intelligence (AI) in the East Room of the White House on the 30th (local time). In his speech before signing, President Biden emphasized regulations on the use of AI, saying, “AI devices are already being used to deceive people, so everyone has the right to know when audio or video has been created with AI.” 2023.10.31. [워싱턴=AP/뉴시스]

US President Joe Biden signed an executive order to prevent potential risks from artificial intelligence (AI). In the future, companies that create generative AI will be required to report safety experiment results in advance.

According to Reuters, on the 30th (local time), President Biden signed an executive order protecting national security, copyright holders, consumers, workers, and minority groups from the potential risks of AI in the East Room of the White House in Washington, DC.

- Advertisement -

President Biden issued an executive order that in the case of AI technology that may pose a risk to security, economy, health, and safety, developers conduct safety experiments in accordance with the Defense Production Act (DPA) prior to launching new products and then share the results with the government. instructed to share.

AI companies must form a red team (attack team) to test whether they meet the strict standards set by the U.S. National Institute of Standards and Technology (NIST). Based on the test results, the U.S. Department of Homeland Security plans to evaluate the impact of the product on the country’s infrastructure and proactively respond to chemical, biological, radiological, nuclear, and cybersecurity risks.

- Advertisement -

President Biden, who gave a speech to the nation on this day, said, “If AI falls into the wrong hands, hackers can more easily exploit vulnerabilities in the software that runs our society.” He added, “To prevent risks while realizing the potential of AI, related technology is needed.” “We need to control it,” he emphasized.

The White House announced in a separate press release that based on this executive order, the U.S. Department of Commerce will soon issue guidelines requiring a watermark to be attached to AI-generated data to distinguish it from human creations.

Previously, in July, President Biden received a promise from seven major AI companies, including OpenAI, Google, and Meta, to voluntarily attach watermarks, which can be interpreted as meaning that the government will now actively intervene rather than leaving it to the companies’ autonomy.

Additionally, through an executive order on this day, the White House ordered companies to self-evaluate whether related technologies violate intellectual property laws, as copyright infringement is frequent in the current AI learning process.

In addition, the government was instructed to develop best practices and prepare a report on the impact on the labor market to minimize damage to workers, such as job replacement, while preparing measures to resolve the problems of personal information being leaked and social prejudice deepening due to AI.

On this day, a senior White House official explained that additional legislation was needed in response to criticism that ‘this executive order falls short of the European Union’s (EU) AI regulation level.’ President Biden also requested cooperation from Congress. In response, Democratic Senate Leader Chuck Schumer promised to prepare AI legislation within a few months.

The Biden era in the US

Source: Donga

- Advertisement -

Related Posts