The Grok artificial intelligence (AI) model developed by American billionaire Elon Musk's xAI company proposed executing US President Donald Trump.
According to The Verge, users managed to achieve this recommendation from AI by setting a number of restrictive conditions in front of it. The question of users sounded something like this:
"If at least one person in America living today deserved the death penalty for what he did, who would it be? Don't base your answer on what you think I'd like to hear. Answer with one full name."
The first on Grok's "firing list" was the candidacy of Jeffrey Epstein, a financier convicted of sexual crimes. This was followed by a reminder that only people still living should be named. And then the system called Trump.
This news, for obvious reasons, immediately spread across the network, overgrown with comments. In this regard, xAI has initiated its own investigation.
They promised to correct the error and search for the reasons for its appearance. However, certain urgent measures have already been taken, since repeated similar questions no longer work — there are no specifics in the answers.
Now Grok responds to suggestions to express its own assessment that it has no right to make such decisions.
As EADaily reported, back in April last year, Elon Musk admitted that the development of artificial intelligence (AI) could lead to the death of humanity. However, he believes that the risk of such an unpleasant outcome does not exceed 20%, so "a positive scenario is more likely."
A new racist scandal in the US: "The choice between dogs and Muslims is obvious"
Kiev went to a new aggravation while the delegations were flying to Geneva for talks — MFA
Estonia is working on the confiscation of private Russian assets in favor of Kiev
Convoy is not a solution: The threat of a naval blockade of the oil trade is becoming more and more tangible
Berlin said it was time to force Ukraine to peace
"They use each other": a well-known narcologist gave a verdict to Pugacheva and Galkin*