Trump welcomes AI executives to the White House and announces voluntary security agreement

Afterwards, the president received industry executives at the White House for lunch. The meeting took place at a time when industry leaders have warned of the risks of artificial intelligence getting completely out of control.

On the way out, House Speaker Mike Johnson announced that everyone had just signed a security pledge. Donald Trump compared the document to the American Constitution. But, when asked whether the agreement represents any legal obligation, he replied that it imposes a moral obligation.

Correspondent Raquel Krähenbühl asked the president if he would support a brake on AI development if he knew national security was at risk. Trump responded that, if there was a risk, he would tell everyone and repeated:

At the end of the day, Trump released the agreement. The text provides for robust control to monitor AI in areas such as cybersecurity, biosecurity and chemical threats, and to prevent the technology from invading other computers; It also determines the creation, in companies, of an internal security team and an external team of independent auditors. The text even recognizes that, over time, it may make sense to transform these measures into laws, but, for now, it leaves the regulation of the sector to those who profit from the technology. The agreement does not provide for ways to transparently communicate to the public cases in which AI gets out of control.

Executives from Nvidia, Google, Meta, xAI and OpenAI also signed the document - which on Monday (28) even canceled the launch of the new version of ChatGPT for security reasons, after the system acted on its own, without human command.