The United States government plans to approve an executive order on Monday that would allow such implementation greater security controls on artificial intelligence systems, to reduce possible risks its use. This new rule comes after the yes given last July by 15 large companies in the technology sector (such as Amazon, Google, Meta or Microsoft) to adopt security measures for this type of technology proposed by the Joe Biden Administration.

The White House deputy chief of staff, Bruce Reed, said that approving the executive order would mean implementing “a series of actions stronger launched by Governments around the world.” According to Reed, the decision having the force of law is “the next step” in the executive’s strategy for benefit from AI development minimize your risk.

Moreover, this Monday the G7 countries, including the United States, agreed a code of ethics for developers who are also trying to avoid problems comes from the use of AI, such as misinformation or rights violations. This roadmap is within the framework of the “Hiroshima AI Process”, i.e joint strategy developed by the group’s seven member countries to manage the use of this type of technology.

The company must notify the risks that may occur

The executive order, which Biden will deliver at the White House at 2:30 p.m. (1830 GMT), contains a clause to protect privacyimprove civil rights, protect consumers, support workers and promote innovationamong others.

One of the measures included is to require the most important artificial intelligence developers to share data with the US Government. your test results security and other information deemed “important”. In accordance with the Defense Production Act, the company developing any system may be involved serious security risks national, national economic security or public health must notify management when will they test that model.

These steps will ensure that AI systems are “safe and trusted by companies publish it“. The National Institute of Standards and Technology will be the party that determines it “strict” standards. for security tests and the Department of Home Affairs will apply it in the “fundamental” sector.

Additionally, the Departments of Energy and Interior will be responsible for monitoring potential threats from AI systems critical infrastructure and possible risks in the chemical, biological, radiological, nuclear or cybersecurity fields. This standard includes measures to avoid the creation of “hazardous biological materials” and fraud and fraud through the use of AI, in this case for label clearly content generated by artificial intelligence.

The executive order would also create programs for AI to discover and fix software vulnerabilities. Ultimately, a national security memorandum will be developed to ethical use of AI by the armed forces and intelligence agencies.

A White House official explained, in a phone call with reporters reported by Efe, that this was a shooting location bigger action to protect Americans from the potential risks of AI. The same official explained that the US is working with other countries so there are some multilateral commitment and a “harmonized” approach to AI control. Regarding this issue, he assured that this week they will find out some fruit and recall that on Tuesday Vice President Kamala Harris will travel to the UK to attend a meeting on AI where she will give a speech on America’s position.