This June 2nd, President Donald Trump signed an executive order seeking to give his Government access to artificial intelligence (AI) models to evaluate cybersecurity risks.
According to CNN, the executive order requires that AI companies voluntarily share their new models that they believe have advanced cyber capabilities, 30 days before they are provided to business partners.
The outlet claims that Donald Trump was initially expected to signed said executive order two weeks ago, but it was postponed.
A White House spokesperson said this executive order seeks to “reflect Trump’s common sense approach to support the AI industry and thus balance innovation and security“.
“It seeks to consolidate the continued global dominance of United States in artificial intelligence and cybersecurity“said the White House.
The executive order, CNN states, also asks national security agencies to improve their defense systems regarding cybersecurity through a so-called “clearing house”.
It is also highlighted within the order that “nothing related to the voluntary publication of AI models”should be interpreted as authorizing the creation of a mandatory government license or permit requirement“says CNN.
The media adds that, so that the Trump Government developed the executive order, had the support of large AI companies.
One of them, the media points out, is Anthropic, a company that was recently included in the Pentagon’s “black list”, since it was considered “a risk to the supply chain.”
In fact, Trump’s executive order was initially going to take a more passive stance, but it wasn’t until Anthropic introduced its AI model called Mythos, which, according to the company, “can exploit cybersecurity vulnerabilities at an unprecedented rate“.