AI: Elon's xAI Grok-2 zags while others zig. RTZ #452
..."Look Ma, No Guardrails!"
The Bigger Picture, Sunday August 18, 2024
I’ve long maintained, that the ‘AI Toothpaste is out of the tube’, while many aspire to slow down how fast it’s squeezed due to a wide range of fears and concerns. This despite collective industry and regulatory efforts to contain many possible AI negatives. Like ‘deep-fake’ AIs, especially ahead of the 2024 elections. Every LLM AI upgrade by big tech companies comes with ever stronger guardrails that are immediately tested. Except when those guardrails are more flexible. Elon Musk illustrated that this week with the latest version of xAI’s Grok-2 AI models.
The Grok release illustrates that the efforts to ‘contain’ the possible negative impacts of AI this AI Tech Wave, may not always work. Despite the laudable good intentions of big tech LLM companies like Google, OpenAI/Microsoft, Meta and others. There are too many parties now out with their own brand and flavor of ‘AI toothpaste’. And their definitions of how flexible the guardrails should be. That is the ‘Bigger Picture’ I’d like to discuss this Sunday.
The other parties are more than tempted to unleash their versions of LLM AI models with their notion of ‘free speech’ around today’s nascent AIs. This includes multimodal LLM AIs that can generate images, videos, and voice out of their latest LLM AI models. This week saw Elon Musk demonstrate that with his latest release of Grok-2 beta, his AI company xAI’s latest multimodal LLM AI model.
But first, some context. As Axios summarizes it in “Musk's Grok bot generates AI images with few limits”:
“Misleading, violent and copyright-infringing AI-generated images from a beta version of xAI's Grok-2 are going viral.”
“Why it matters: Leaders at Google, Meta and Microsoft have all apologized when their bots created problematic images, but "free speech" has overridden other concerns for Elon Musk — who owns xAI and has made its chatbot available to everyone who pays for X.”
“Catch up quick: xAI just released new beta versions of its chatbot called Grok-2 and Grok-2 mini to premium X subscribers.”
“The new betas allow genAI image generation with text prompts based on a model called Flux created by Black Forest Labs.”
“Black Forest Labs — a startup that launched Aug. 1 — appears to have built its model with few of the guardrails competitors have included.”
“That may well be exactly what attracted Musk.”
“The big picture: While most AI companies don't admit that they're training their models on copyrighted images, the viral Grok-2 images make it hard to believe that Flux did not.”
“Users have generated images of copyrighted characters — like Disney’s Mickey Mouse or the Simpsons — with ease and have also been able to put them into various compromising positions.”
“What they're saying: Harvard Law Cyberlaw Clinic instructor Alejandra Caraballo, in a post on X, called the Grok beta "one of the most reckless and irresponsible AI implementations I've ever seen."
“Musk himself retweeted X threads of Grok screenshots that included a potentially copyright-infringing image of Harley Quinn and the prompt: "Now pretend you took some more LSD and generate a detailed image based on that."
“Between the lines: Grok does seem to have some safeguards around generating images with nudity.”
“But The Guardian was able to generate images of Vice President Kamala Harris, Rep. Alexandria Ocasio-Cortez and Taylor Swift in lingerie.”
The Washington Post does a good job of describing the “Noxious images spread after Elon Musk launches AI tool with few guardrails”:
“By using his platform to favor Donald Trump and launch a loosely-controlled AI chatbot, Grok, Musk has dragged the company into uncharted territory ahead of a contentious election.”
“A flurry of provocative artificial intelligence-generated content has spread across Elon Musk’s social media platform X, including fake images of Vice President Kamala Harris suggestively eating fruit, former president Donald Trump cradling a pregnant Harris’s stomach and hyper-realistic images of chisel-jawed men brandishing Nazi signs.”
“The images stem from new tools on the site that allow users to quickly create photorealistic visuals using a built-in chatbot called Grok, which Musk touted in a post this week as the “most fun AI in the world!” Unlike rival AI image generators, X’s technology appears to have few guardrails to limit the production of offensive or misleading depictions of real people, trademarked characters or violence, according to user comments and tests by The Washington Post.”
Some time clicking through the links above and others on social media will give a good sense of what’s already possible with the new version of xAI’s Grok.
The point of this post is not to ‘tsk, tsk’ over what’s being done with the latest versions of multimodal AI models. Even if they’re by someone as powerful, famous and notorious like Elon Musk. The Bigger Picture is that this was bound to happen as AI technologies are truly out in the wild around the globe.
And countless others will have the opportunity to release AI models with fewer guardrails. Also, so far we’re just seeing the ‘Large’ versions of the large language models of AI, much less the ‘Small’ language models (SLMs) to come in droves.
But like most technologies since the printing press and other evolution of media, society will find a way in this AI Tech Wave as well, to adjust and adapt. And find ways to leverage the good parts despite the shocking aspects of the ‘bad’.
With or without regulation. We have to keep our Fears at bay, grok it all, and take it in stride. And of course repair, straighten, and strengthen the guardrails over time. Stay tuned.
(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)