Grok AI accused of making obscene pictures
Digital Desk, New Delhi. Governments around the world have expressed concern and criticized the creation of indecent images of women and children by Grok, the AI chatbot of Elon Musk's Platform
Britain's top technology official demanded on Tuesday that Ax take immediate action. The EU's executive branch has condemned Grok while officials and regulators in India, France, Malaysia and Brazil have condemned the platform and called for an investigation.
Growing concern from individual countries points to the sinister potential of apps that use AI to generate sexual deepfake images.
Grok AI accused of making obscene pictures
The problem came to light after the launch of Grok Imagine last year. It allows users to create videos and pictures by typing text prompts. It includes a "Spicy Mode", which can create adult content.
This became more dangerous late last month, when Grok began accepting large numbers of user requests to modify photos posted by others.
Nonprofit group AI Forensics said it analyzed 20,000 photos taken by Grok between Dec. 25 and Jan. 1 and found that 2 percent depicted a person 18 or younger in a bikini or transparent clothing.
X does not deny that such material generated through Grok exists. It claimed in a post that it takes action against illegal content, including child sexual abuse material. However, the list of countries demanding a ban on such content on X is growing.
The Indian government on Friday issued an ultimatum to X, asking it to remove all illegal content and take action against such users. The country's Ministry of Electronics and Information Technology ordered the company to review Grok's "technical and governance framework" and file a report on the action taken.
The ministry accused Grok of 'gross misuse' of AI and serious failures of safeguards and enforcement by allowing women to be photographed or shared in a 'derogatory or obscene' manner to defame them in an indecent manner.
Warned that the company would face legal problems if the 72-hour deadline was not followed. This deadline has passed.
(With inputs from news agency AP)