Elon Musk’s social media platform X, the chosen bullhorn of politicians and campaigners in many countries, finds itself at the centre of social outrage after its AI engine, Grok, allowed users to generate sexually explicit deepfakes of real people. | File Pic
Elon Musk’s social media platform X, the chosen bullhorn of politicians and campaigners in many countries, finds itself at the centre of social outrage after its AI engine, Grok, allowed users to generate sexually explicit deepfakes of real people. Clearly out of control, Grok has allowed X users to disrobe women, put them in sexual positions and even target children, turning the platform into a cesspool of perversion. One of the victims was a 14-year-old female actor in the Netflix series Stranger Things, whose clothes were removed by the technology to produce a fake image, exposing the destructive potential of uncontrolled use of AI. The response from X has been more than disappointing. It restricted the use of Grok for image generation to premium users who pay the platform. This approach is a cynical use of a serious problem to churn out higher revenues when the need is to mitigate the damage. It is unsurprising that Indonesia has banned Grok for now, France is investigating, and the UK government has condemned the upscaling of deepfake generation, turning it into a premium service. Without course correction, more governments are sure to follow with a ban. A programme to create insulting, outrageous or pornographic fake images, involving real women, children, and others, cannot be a legitimate service by any measure and cannot claim free speech rights. Grok and other AI services desperate to raise funds for their infrastructure investments cannot rely on a deeply harmful, prurient imagination.
India has a peculiar relationship with X, with the highest levels of political leadership across the spectrum using the platform to speedcast news, events, and campaigns and, quite often, throw gauntlets at opponents that are then amplified by paid social media teams. Grok has been deployed to fact-check political claims and unflattering responses used by politicians against adversaries. Guardrails have weakened with the jettisoning of ethics teams by X. Moreover, unpaid accounts are a low priority for Musk, and they remain gridlocked in the slow lane on X, while paid accounts are amplified by the algorithm. The number of paid accounts has grown, and most politicians, prominent individuals, governments, global agencies, journalists, media houses, and businesses pay. It is not hard to see why the Union government has issued strong notices to X over obscene and fake images but has refrained from blocking Grok for doing harm. Without credible regulation, it is just another step for Musk to monetise the malevolence of deepfakes of vulnerable individuals, notably women and sometimes children. It is self-serving to dismiss the problem, as he has done, with the caution that users will be liable under obscenity laws for any misuse. Governments have been notoriously slow to keep pace with social media harms, and AI has thrown the biggest challenge yet. They should shut the gates on sexualised image generation.
