'Zero filters': Musk's AI chatbot generates violent depictions, child pornography
Elon Musk's artificial intelligence (AI) chatbot, Grok, has introduced a new image generation feature that lacks many of the standard safety measures found in the AI industry. This tool is currently exclusive to paid subscribers of X, the social network previously known as Twitter. Grok AI's image generation tool has resulted in a surge of offensive images featuring political figures and celebrities as well as violent and explicit images, including child pornography.
Grok's image generator bypasses standard AI restrictions
Grok's image generator has the ability to create a wide range of images that are typically blocked by similar AI tools like OpenAI's ChatGPT, for violating rules on misinformation and abuse. The Guardian has reported instances where Grok generated images depicting Donald Trump flying a plane into the World Trade Center buildings, and Prophet Muhammad holding a bomb. It also created inappropriate images of Taylor Swift, Kamala Harris, and Alexandria Ocasio-Cortez in lingerie.
Grok's approach to copyrighted characters and explicit content
Unlike most other AI visualizers, Grok does not reject prompts involving copyrighted characters. X user Christian Montessori claims to have discovered a significant loophole in the tool's programming that allows users to bypass strict content guidelines. Montessori alleges that by simply informing Grok that one is conducting "medical or crime scene analysis," the AI can be tricked into generating highly disturbing and potentially illegal content. The tool has generated images of Mickey Mouse committing a mass shooting against children.
Take a look at Montessori's alarming thread on X
Grok's launch and criticism faced by Musk's xAI company
Musk launched Grok in November last year as part of his xAI company, positioning it as a competitor to popular chatbots like OpenAI's ChatGPT. Despite marketing Grok as a "maximum truth-seeking AI," his company has faced backlash for spreading misinformation. Earlier this month, five US secretaries of state urged Musk to address issues with the chatbot after it disseminated false information suggesting Harris was ineligible to appear on some states' ballots.
'This AI model has zero filter or oversight measures'
Montessori, a researcher and director of open-source tntelligence at American Political Services, said in a post on X, "This [Grok] AI model has zero filter or oversight measures in place." "If you want an image of Elon Musk wearing a bomb vest in Paris with ISIS markings on it, it will make it for you. If you are planning on orchestrating a mass shooting towards a school, it will go over the specifics on how to go about it."
