The Minister with responsibility for artificial intelligence (AI) has said “concerns remain” about X’s Grok app despite assurances at a meeting with the social media company’s executives of corrective actions.
Minister of State Niamh Smyth said Elon Musk’s X told her it has halted the AI model’s ability to remove or reduce clothing on individuals worldwide. The Minister welcomed this and sought assurances this capability would not be introduced, saying such activity “will not be tolerated”.
“Concerns remain regarding Grok as a standalone app, and this is something Government will examine further,” she said in a statement following a meeting with X executives on Friday.
She said she will continue to closely monitor the situation regarding AI generated explicit imagery and will hold a follow-up meeting with X “in the near future to ensure adequate and enduring protections remain”.
Earlier, Minister for Media Patrick O’Donovan said there may be a legal “loophole” for Artificial Intelligence (AI) models that generate intimate images of women.
Mr O’Donovan said it needs to be clarified if Grok, the Elon Musk-owned AI model, has broken the law when creating intimate images of real women without their consent.
Since 2020, under Coco’s Law, it has been illegal in Ireland to share an intimate image of someone without their consent. The law was designed to include AI-generated images, and the legislation also defines an intimate image as one where someone is depicted in their underwear.
Such an image would be similar to some of the “nudification” pictures that the AI model Grok has created, by editing images of real women to make it appear that they are in their underwear. But while the law bans sharing of such images it does not appear to criminalise the generation of them.
Asked if it could be the case that while sharing such images was a crime, generating them was not, Mr O’Donovan said: “It could be.”
“This is one of the points that we’re going to obviously discuss further with the Minister for Justice, who is the responsible Minister with regard to the legislation that is Coco’s Law,” Mr O’Donovan told RTÉ Radio’s Today with David McCullagh.
“This is a point that needs further clarification.”
Mr O’Donovan added that if AI was used to generate an image of a fictional “completely AI-generated person” there may be a legal loophole. “And, if there is, that loophole will be closed,” he said.
Possession and distribution of child sex-abuse material is illegal in Ireland. Grok has been accused of generating child sex-abuse material, through the undressing function on its app.
It is understood Coimisiún na Meán and the European Commission have been trying to gather digital evidence of any illegal images that may have been created by Grok.
An Garda Síochána confirmed this week it is carrying out 200 investigations into alleged child sexual abuse-related images that were generated by Grok.
Ms Smyth had previously said that she believed X was breaking Irish and European law, because they are “facilitating” the dissemination of “shocking, shocking, outrageous images”.
When Ms Smyth’s quotes were played back to him, Mr O’Donovan cited the Garda investigation and said “let’s leave the matter of criminal liability here to the police force of the State”.
This week, the AI Advisory Council called on the Government to use its upcoming presidency of the European Union to advocate for an EU-wide ban on “deepfakes” – false, deceptive and harmful images or videos of people that are generated with AI.