Elon Musk’s AI company, xAI, has just rolled out an early preview of Grok-2, and it’s causing quite a stir.
I found out why when speaking with Marketing AI Institute founder and CEO Paul Roetzer on Episode 111 of The Artificial Intelligence Show.
Image Generation Controversy
xAI claims that an early version of Grok-2 was able to outperform both GPT-4 Turbo and Claude 3.5 Sonnet on the LMSYS.org chatbot leaderboard.
But it’s not the improved reasoning, code, or language capabilities of the models that have everyone talking…
Instead, it’s the image generation capabilities.
Grok-2 allows you to generate images using text prompts by integrating with a powerful new image generation model called Flux.
And it appears that Musk and xAI have included very few, if any, guardrails on the images you can generate and post to X. There appear to be no real restrictions on prompts involving real people and no watermarks on the generated images.
This has quickly led to users creating profane images of real people (including political figures) and images that very clearly use copyrighted material from brands like Disney.
“It will create anything,” says Roetzer.
The Disney Factor: A Potential Legal Showdown
One particularly interesting aspect of this story involves Disney. Roetzer tested Grok-2’s image generation capabilities with Disney characters and found the results eerily accurate:
“My first test was Disney characters as an experiment and it did well. I did Mickey Mouse and Goofy riding in a Tesla and it was as though Disney created it,” he says.
This capability, combined with Musk’s recent public clash with Disney CEO Bob Iger over advertising on X, leads Roetzer to speculate that Musk might be baiting Disney into a legal confrontation.
“It’s what Musk does. He pushes the boundaries and invites lawsuits… I think it keeps him entertained and motivated to do what he’s doing,” says Roetzer.
The Bigger Picture: xAI’s Rapid Rise
To understand the significance of this development, it’s crucial to consider XAI’s rapid ascent in the AI world, says Roetzer.
The company was founded in July 2023, released Grok-1 in November 2023, and raised a $6 billion Series B funding round at a rumored $24 billion valuation.
In other words, the company is moving at breakneck speed to develop and deploy AI. And, in typical Musk fashion, it doesn’t appear to mind courting controversy along the way.
And there’s plenty of that.
The unrestricted nature of Grok 2’s image generation capabilities raises serious legal and ethical questions around copyright infringement, trademark violations, and potential defamation issues.
As Roetzer points out:
“Elon’s one of the few people who’s got the money, and the personality to sort of take this kind of stuff on in the name of ‘free speech.’”