It is not yet clear if the broad backlash against artificial intelligence (AI) models which allow the generation of intimate images of adults and child sex abuse material marks a turning point in how people look to their governments and regulators to rein in Big Tech, but it may be.
Regulating the social media behemoths is hugely complex and painstaking. To date it has been largely left to regulators and the higher courts to pursue and adjudicate on matters such as antitrust breaches or non-compliance with digital rules.
This can take years.
The realisation that Elon Musk’s AI chatbot, Grok, allowed the generation of so-called “nudification” deepfake images of adults and children, caught the attention of the public and they demanded swift action from governments.
READ MORE
Creating these non-consensual images was possible because the Grok AI product is designed to have fewer so-called “guardrails” than competitor AI products, based on Musk’s belief in a “maximally truth-seeking” model.
Some of the public reaction flowed from Musk’s initial response, which involved crying emojis, decrying “any excuse for censorship”, before limiting access to the feature to paying subscribers.
In keeping with the “we are just a platform” mindset - shared with other social media firms - X also sought to place the responsibility for the creation of such images on the users saying, “anyone using or prompting Grok to make illegal content will suffer the same consequences as if they upload illegal content.”
Had Musk apologised and pledged to address the issue quickly and decisively, public interest would have likely quickly moved on.
But he didn’t.
National governments in the UK, France, India and here pledged to act and X faced threats of fines and bans. The European Commission ordered X to retain internal documents relating to Grok until the end of the year.
Apple and Alphabet’s Google faced calls to remove X and its AI tool Grok from their app stores due to the spread of the sexualised images on the social media platform.
Law enforcement agencies got involved. Gardaí said they were investigating 200 reports involving content on the social media platform X that is allegedly child sexual abuse material.
On Thursday X, which has its European headquarters in Dublin, said its Grok AI function would no longer allow users to manipulate photos of people to appear in revealing clothing in places where such actions were illegal. The Irish Government welcomed X’s “corrective action”.
Following a meeting with X on Friday, Minister of State Niamh Smyth said the social platform had told her it has halted the AI model’s ability to remove or reduce clothing on individuals worldwide. She said it had also conceded that it had hosted content that is illegal in Ireland.
In the midst of an almost overwhelmingly swift and complex AI revolution, the Grok case was simple for the public to grasp, and led to widespread opposition and loud cries for an immediate response.
Following Musk’s apparent climbdown it will be interesting to see if the now spurred public, politicians, law enforcement and regulators sustain their interest in the emerging capabilities of AI models and demand their government’s seek to impose tighter regulations.
The issue has prompted much coverage in The Irish Times. This weekend Patrick Freyne examines the descent of Twitter from a social media site which re-engineered global discourse to the Musk-owned X “riddled with bigotry where you couldn’t be sure you wouldn’t accidentally scroll past illegal content.”
Finn McRedmond notes that the fact that this debate over deepfake sexually explicit images of women, and in some instances, children, would have been incomprehensible to the average reader just a year ago speaks to the vertiginous pace at which the internet develops. Despite this, she says calls to ban X are misguided.
“Contained within the Grok and X saga in Ireland and the UK, I see two countries without a firm or clear grasp of their own principles. Or two governments afraid to make the unfashionable argument that spheres of public conversation are sacred and should not be subject to nationwide bans."
Meanwhile, columnist Mark O’Connell recalls meeting contractors who worked in AI safety whose job it was to ensure that the company’s Large Language Model software could not be used to create content that was either illegal or against the company’s internal ethical guidelines. “The company they contracted for, I need hardly tell you, was not X”.
Inside Politics
From tomorrow our Inside Politics newsletter will now issue five days a week – up from the current three, on the weeks the Dáil is sitting. Our new editions on Mondays and Fridays will allow our political team to deliver even deeper coverage of the Irish political world direct to your mailnbox. You can sign-up to the digest here.
We value your views. Please feel free to send comments, feedback or suggestions for topics you would like to see covered to feedback@irishtimes.com.
