The President of the United States, Joe Biden, plans to approve an executive order this Monday to allow greater control over the security of artificial intelligence systems and reduce the risks of this technology.
According to information provided by the White House, this is “the most powerful set of actions launched by a government anywhere in the world.”
This was stated by the deputy chief of staff, Bruce Reed, who pointed out that this decree with the force of law is the “next step” in an “aggressive” strategy to act on all fronts, “taking advantage of the benefits of AI and mitigating its risks.” .
The rule comes after 15 of the country’s leading technology companies, including Amazon, Google, Meta, and Microsoft, agreed to adopt security measures in the development of AI as proposed by the Biden Administration.
The executive order, which Biden will present at the White House at 2:30 pm (6:30 pm GMT), includes clauses to protect privacy, improve civil rights, protect consumers, support workers, promote innovation, etc.
Among the measures it includes is a request that the most important developers of artificial intelligence share with the US government the results of their security tests and other information that is considered “important.”
Under the Defense Production Act, companies developing any system that could pose a serious risk to national security, national economic security, or public health must notify the administration if they plan to test it. model.
These measures will ensure that artificial intelligence systems are “secure and reliable before they are introduced to companies.”
The National Institute of Standards and Technology will establish “strict” standards for security testing. And the Department of the Interior will apply it to “fundamental” sectors.
In addition, the Departments of Energy and Interior will be in charge of monitoring possible threats from AI systems to critical infrastructure and the risks they may pose in the chemical, biological, radiological, nuclear, or cybersecurity fields.
The standard includes measures to try to prevent both the production of “dangerous biological materials” and fraud and deception through the use of AI, in this case the clear label of the content generated by artificial intelligence.
The executive order will also establish a program for AI to discover and fix software vulnerabilities. Finally, a national security memorandum will be drawn up for the ethical use of AI in the armed forces and intelligence services.
The executive order also includes measures to protect privacy, improve civil rights, protect consumers, support workers, and promote innovation, among others.
A White House official explained at a press conference that this is the largest set of measures to protect Americans from the potential dangers of AI.
The same official explained that the US is working with other countries to have multilateral commitments and a “harmonized” approach to controlling AI.
On this issue, he assured that some fruits will be known this week and recalled that on Tuesday, the vice president, Kamala Harris, will travel to the United Kingdom for an AI meeting where she will give a speech on the position of America.