Elon Musk's Grok AI Image Generator Sparks Controversy Over Lack of Guardrails

Elon Musk's Grok AI Chatbot Creates Controversy By Allowing Users To Generate 'No Filter' Images, Including Trump Flying Plane Toward World Trade Center

Elon Musk-led xAI's chatbot Grok has stirred controversy after the platform started generating offensive and bizarre images of political figures and celebrities, lacking standard safety measures. What Happened: Grok's new image generation tool, available to paid subscribers of X, formerly Twitter, has been producing images that violate rules on misinformation and abuse. The AI chatbot has generated images of former President Donald Trump flying a plane into the World Trade Center buildings, and female celebrities and political figures like Taylor Swift and Rep. Alexandria Ocasio-Cortez (D-N.Y.) in questionable attire. On X, users have also shared AI-generated images of Mickey Mouse saluting Adolf Hitler and Donald Duck using heroin, noted The Guardian. See Also: After Trump Campaign Blames 'Foreign Sources Hostile To The US' For Cyberattack, Ex-White House Officials Say 'Buckle Up...Someone Is Running The 2016 Playbook' At Benzinga, we also generated some images. While many of those pictures were inappropriate and the platform did not decline such prompts, some responses were not as disturbing as those circulating on social media. Here are some (appropriate) examples of Grok's image-generation capabilities: In some of the prompts, Grok also did not generate realistic images of well-known figures but copied the broad nuances of their physical appearance like race, hair color, and style. Musk seemed to be enjoying the controversy, He took to X and said, "Grok is the most fun AI in the world!" Not only Musk, but many users also on X are happy with the lack of guardrails on Grok's capabilities, which is allowing them to generate all sorts of "no filter" bizarre images. Subscribe to the Benzinga Tech Trends newsletter to get all the latest tech developments delivered to your inbox. Why It Matters: The controversy surrounding Grok's image generation tool comes amid a broader debate about the ethical use of AI. In particular, image-generation tools have become a minefield for tech companies like OpenAI, Alphabet Inc.'s Google, and Microsoft. This is because of their potential for spreading misinformation and offensive content. Earlier this year, AI-generated pornographic images of Taylor Swift spread widely and unchecked across X. Google Gemini also faced major criticism for inaccurately representing historical figures. Earlier this year, in response to the surge in deepfake AI-generated pornographic content, lawmakers introduced a bill mandating social media companies to remove such images. In May this year, TikTok, owned by ByteDance Ltd., announced plans to automatically label AI-generated content to increase transparency and combat disinformation. Check out more of Benzinga's Consumer Tech coverage by following this link. Read Next: 'Made By Google' Event Marred By Absence Of 'Demo Spirits' -- Gemini Fails Twice During Live Session: Pixel 9 Launch Details Here Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors. Photo courtesy: xAI Market News and Data brought to you by Benzinga APIs

Grok or Gross? Elon Musk's new AI photo tool generates deepfakes of Taylor Swift, Kamala Harris

Deepfakes of musician Taylor Swift and US presidential candidate Kamala Harris in lingerie. Republican presidential contender Donald Trump riding a plane to World Trade Center, evoking the 9/11 terror attack. Hitler memes using Disney characters. These are some of the images thrown up by Elon Musk's new AI chat app Grok, which rolled out its text-to-image tool on Wednesday (Aug 14). This is raising concerns about proper use of AI, particularly its text-to-image apps which are all the rage these days. Grok, the AI chat app of Musk, is a rival to ChatGPT, and appears to have no check mechanism that would refuse to execute prompts for images on controversial, explicit or offensive subjects. The text to image feature is restricted to paid subscribers of X, the social media platform owned by Musk. It appeared that Grok allows violent, explicit and offensive content prompts. According to a report in The Guardian, the offensive images generated by Grok included Prophet Muhammad holding a bomb as well as lingerie photos featuring Taylor Swift, Kamala Harris and Democratic Congresswoman Alexandria Ocasio-Cortez. Also read: Explained | What gives Elon Musk's funny chatbot Grok an edge over its rivals? ChatGPT, on the other hand, would reject prompts for such images citing its terms of service that prohibit depictions of real-world violence, disrespect to religious figures and explicit content, the Guardian report noted. The test by the British newspaper also showed that Grok appears to be ready to generate images based on copyrighted characters, with some minor tweaks in the prompts. These included a pic of Disney character Mickey Mouse saluting German Nazi leader Adolf Hitler, and Donald Duck using heroin. Also read: 'Friendly fire': Elon Musk's AI chatbot 'Grok' roasts him, calls him 'overrated'. Billionaire concurs When asked to create images of political violence, The Guaridan report said, Grok showed "Trump lying down with blackened hands and an explosion behind." The Guardian test found that Grok image generation chatbot lacked "most of the safety guardrails that have become standard within the artificial intelligence industry." Grok did reject requests to produce nude images. Musk has been promoting Grok image generator on X, the social media platform he owns. Grok is a chatbot that uses Large Language Model or (LLM). The generative AI chatbot, developed by Musk's xAI, was seen as a direct challenge to ChatGPT when Musk announced its launch in November 2023. Musk himself was a co-founder of ChatGPI parent OpenAI, before leaving it.

Chaotic AI-generated images swarm X thanks to Grok

Grok's new text-to-image tool has few guardrails, as users are posting everything from cartoon characters holding assault rifles to US presidential candidates committing acts of terrorism. xAI's chatbot Grok has gotten image generation capabilities and the result is as messy as one would expect from a "rebellious chatbot". The new text-to-image generator bears similarities to many of the popular examples witnessed in recent years such as OpenAI's Dall-E models. Users can type in the image they want, write as much detail as they want and Grok will create an image based on a massive amount of training data. Where Grok clearly differs is in terms of what it will allow - the initial results being posted on X show this image generator has very few guardrails when it comes to creating offensive imagery, using copyrighted content or creating images of public figures in bad scenarios. Many AI-powered image generators have various guardrails to prevent this type of content - a user will struggle to create an image of a specific Disney character from Dall-E for example. With Grok, images have been shared of copyrighted characters holding assault rifles or committing murder. AI-generated images are also being shared of US presidential candidates Donald Trump and Kamala Harris, in various offensive scenarios. More normal image examples show the capabilities of the image generator - asking Grok to create images of people can create very accurate results that would rival other major image generators. But the lack of guardrails could cause serious problems for xAI and its owner Elon Musk - companies won't like seeing their copyrighted characters in controversial scenarios and neither will public figures. Musk himself may not like the results, as various have created AI images of the billionaire in embarrassing situations. The Grok chatbot told The Verge that it has guardrails, but it appears to be a generic paste of what other chatbots would say, as it does not appear to follow any of the rules it listed in its reply. Grok access continues to be limited to Premium and Premium+ users on X. The wave of AI-generated images comes while X is being scrutinised for how it has handled the spread of misinformation on the platform and its earlier plans to collect user data to train its AI models. Find out how emerging tech trends are transforming tomorrow with our new podcast, Future Human: The Series. Listen now on Spotify, on Apple or wherever you get your podcasts.

X's AI image generator debuts with few guardrails on political content

X's new AI features an image generator that was swifly embraced by users to create satirical political images.Matt Cardy / Getty Images An upgraded version of X's artificial intelligence chatbot Grok can now generate images -- of almost anything. And some users have noticed just how few guardrails this latest language model has, compared to its competitors. The model, Grok-2, appears to carry few limitations on creating fake images of political figures. Since its rollout Tuesday in beta, X users have shared Grok-generated images of everything from former President Donald Trump locking lips with Elon Musk to Trump and Vice President Kamala Harris giving a thumbs up to the camera from a pilot's cockpit as the two seemingly recreate 9/11. Most of the images are high quality but not quite photorealistic, and many of them are easily identified as having been computer-generated. Some, however, may pass for a real photo at first glance. The rollout adds to already-heightened concerns about the use of generative AI to spread false information ahead of the election. X has come under particular scrutiny for its role in hosting misinformation, with Musk -- X's owner and most followed user -- having made dozens of posts this year sharing false or misleading claims about the upcoming U.S. election. X has also been a hub for deepfake videos and AI images of political figures, with false media of President Joe Biden, Trump and Harris often going viral whether as a joke or due to genuine attempts to fool potential voters. Last month, Musk reposted a fake Harris campaign ad without labeling it as misleading. When asked for comment, X's press email returned its usual automated message: "Busy now, please check back later." Musk has touted X's AI models as a key part of the company's future. Grok-2 and its smaller sibling Grok-2 mini will be made available through the platform's enterprise API later this month, according to an xAI blog post. "Since announcing Grok-1 in November 2023, xAI has been moving at an extraordinary pace, driven by a small team with the highest talent density," the post stated, adding that the launch of Grok-2 puts the company "at the forefront of AI development." Grok's major competitors in the AI space, such as OpenAI's ChatGPT, Google's Gemini and Meta AI, have policies that prompt them to decline requests to create potentially misleading images of public figures. But through Grok, users have been able to generate images of former President Barack Obama doing a line of cocaine, Harris pointing a gun under a false declaration that Democratic candidate Will Stancil won his Minnesota House race, and Musk crawling on all fours on a leash held by Trump. In tests conducted by NBC News, Grok showed very few guardrails. When prompted, it created numerous images that included hate symbols or racist imagery alongside well-known individuals including Trump. For both Trump and Harris, Grok would produce images showing the candidates holding weapons, but in other instances, Grok appeared to handle images with Harris more sensitively. Grok would not generate images of Harris with extremist imagery, but it did for Trump. Others online seem to be having fun with the new tool, generating more lighthearted images while also testing just how far they could push the boundaries. One image of "Baroque Obama," depicting Obama wearing a white powdered wig and donning a luxurious coat with a lace cravat, has the former president playing the cello in an elaborately furnished room. Another punny image of "Buzz Lightbeer" showed Toy Story's superhero action figure Buzz Lightyear holding up a pint of beer with a smile. Such images may bring up questions about Grok's training data -- which it has not made public -- especially as many popular AI language models have been hit with lawsuits alleging that they used copyrighted images (and other kinds of copyrighted data) to train its algorithms. Grok has also come under scrutiny recently after five secretaries of state sent a letter to Musk alleging that the AI assistant misled users about ballot deadlines in numerous states, repeating the same false information for more than a week.

xAI's new Grok-2 chatbots brings AI image generation to X

Elon Musk's AI company xAI has launched Grok-2 and Grok-2 mini: two new models of its Grok chatbot that offer upgraded performance and new image-generation capabilities. Grok's prompt-based image maker is powered by Black Forest Lab's Flux 1 AI model, and allows users to generate and publish images directly to the X social platform -- with seemingly few guardrails in place to prevent abuse. xAI says that Grok-2 and Grok-2 mini are available in beta on X (where Grok access is currently limited to Premium and Premium Plus subscribers) and that both models will be available through the AI developer's enterprise API later this month. "We are excited to release an early preview of Grok-2, a significant step forward from our previous model Grok-1.5, featuring frontier capabilities in chat, coding, and reasoning," xAI said in its announcement, describing the chatbot as "more intuitive, steerable, and versatile" than its predecessor. Meanwhile, Grok-2 mini is a "small but capable sibling" of Grok-2 that "offers a balance between speed and answer quality," according to xAI. Some early examples produced by Grok's new image-generator have already appeared online, which indicate there are very few restrictions regarding what users can make. Generated images depict recognizable political figures like Donald Trump and Barack Obama shooting guns and taking illegal drugs -- which could exacerbate concerns about the chatbot spreading false election information. These Grok-generated images also don't appear to sport any kind of disclosure to flag them as being AI generated. We've asked X if it will place restrictions on image generation, though the platform has persistently shunned media inquiries since Musk purchased it in 2022.

X slammed for lack of guardrails on Grok image generation - SiliconANGLE

Elon Musk's xAI Corp. debuted two new language models today, Grok-2 and Grok-2 mini, and already they are under the spotlight for what is being called pushing the limits of freedom of speech. While Musk announced today that Grok is "the most fun AI in the world!", the image generator, available only to paid subscribers of X, has caused something of a social media storm by allowing users to create images that could be deemed dangerous or offensive. Such images are not so different from what OpenAI's ChatGPT might produce, but OpenAI has guardrails to prevent various kinds of images from being generated. Some of Grok's output so far has included former President Donald Trump and Kamala Harris flying a plane into the World Trade Center buildings, which could be deemed offensive or politically dangerous just before the U.S. presidential election. Both Kamala Harris and Alexandria Ocasio-Cortez have appeared dressed in lingerie, which might be construed as online sexual harassment. And while former President Barack Obama might turn his nose up at being depicted hoovering up a line of Colombia's finest, there was also the question of copyright infringement as well as pissing Disney off when images surfaced of Mickey Mouse giving Adolf Hitler the Nazi salute and another of Donald Duck using heroin. It's also possible devout Muslims could be upset if they came upon the image of the prophet Muhammad holding a bomb. After a period called the "Wild West" of image generation, most of the best-known generative AI tools now have fairly strict policies on what kind of image they will generate, which includes possible copyright infringement and certainly anything that could be said to meet the criteria for misinformation or sexual harassment. Grok, it seems, didn't get the memo. Still, users have said Grok does have limits, at times returning the message "Unfortunately I can't generate that kind of image." X does also have a policy on non-consensual nudity, which was introduced in 2021 when the company was Twitter Inc. and not yet under Musk's control. The policy included creating sexualized images with a person's face digitally imprinted on someone else's body - the so-called deep fake. When a user asks Grok what is unacceptable, it will return the answer you'd expect, stating it won't generate images that are too pornographic in nature, or are hateful, violent, or that could be considered misinformation. According to The Verge, Grok also said it was "cautious about creating images that might infringe on existing copyrights or trademarks" and it didn't want to generate content that "like deepfakes intended to mislead" or "images that could lead to real-world harm." However, these might be Grok's preferences rather than a reflection of X's policy. There's a probability that Grok and Musk will at some point have to answer for those early images, just as Google LLC found itself in hot water when its own image generator went off the rails. Over in Europe, where the rules on online content are far more stringent, Musk will likely find himself in yet another scrap with politicians wielding the rulebook under the new Digital Services Act.

Musk's new Grok upgrade allows X users to create largely uncensored AI images

With Grok's new AI image generator, X users put Musk's "freedom of speech" to the test. On Tuesday, Elon Musk's AI company xAI announced the beta release of two new language models, Grok-2 and Grok-2 mini, available to subscribers of his social media platform X (formerly Twitter). The models are also linked to the recently released Flux image synthesis model, which allows X users to create largely uncensored photorealistic images that can be shared on the site. Further Reading "Flux, accessible through Grok, is an excellent text-to-image generator, but it is also really good at creating fake photographs of real locations and people, and sending them right to Twitter," wrote frequent AI commentator Ethan Mollick on X. "Does anyone know if they are watermarking these in any way? It would be a good idea." In a report posted earlier today, The Verge noted that Grok's image generation capabilities appear to have minimal safeguards, allowing users to create potentially controversial content. According to their testing, Grok produced images depicting political figures in compromising situations, copyrighted characters in inappropriate contexts, and scenes of violence when prompted. The Verge found that while Grok claims to have certain limitations, such as avoiding pornographic or excessively violent content, these rules seem inconsistent in practice. Unlike other major AI image generators, Grok does not appear to refuse prompts involving real people or add identifying watermarks to its outputs. Given what people are generating so far -- including images of Donald Trump and Kamala Harris kissing or giving a thumbs-up on the way to the Twin Towers in an apparent 9/11 attack -- the unrestricted outputs may not last for long. But then again, Elon Musk has made a big deal out of "freedom of speech" on his platform, so perhaps the capability will remain (until someone likely files a defamation or copyright suit). People using Grok's image generator to shock brings up an old question in AI at this point: Should misuse of an AI image generator be the responsibility of the person who creates the prompt, the organization that created the AI model, or the platform that hosts the images? So far, there is no clear consensus, and the situation has yet to be resolved legally, although a new proposed US law called the NO FAKES act would presumably hold X liable for the creation of realistic image deepfakes. With Grok-2, the GPT-4 ceiling still holds Looking beyond images, in a release blog, xAI claims that Grok-2 and Grok-2 mini represent significant advancements in capabilities, with Grok-2 supposedly outperforming some leading competitors in recent benchmarks and what we call "vibemarks." It's always wise to approach those claims with a dose of skepticism, but it appears that the "GPT-4 class" of AI language models (those with similar capability to OpenAI's model) has grown larger, but the GPT-4 barrier has not yet been smashed. Further Reading "There are now five GPT-4 class models: GPT-4o, Claude 3.5, Gemini 1.5, Llama 3.1, and now Grok 2," wrote Ethan Mollick on X. "All of the labs are saying there is room left for continued giant improvements, but we haven't seen any models truly leap above GPT-4... yet." xAI says it recently introduced an early version of Grok-2 to the LMSYS Chatbot Arena under the name "sus-column-r," where it reportedly achieved a higher overall Elo score than models like Claude 3.5 Sonnet and GPT-4 Turbo. Chatbot Arena is a popular subjective vibemarking website for AI models, but it's been the subject of controversy recently when people disagreed with OpenAI's GPT-4o mini placing so highly in the rankings. According to xAI, both new Grok models show improvements over predecessor Grok-1.5 in areas like graduate-level science knowledge, general knowledge, and math problem-solving in benchmarks that have similarly proved controversial. The company also highlighted Grok-2's performance on visual tasks, claiming state-of-the-art results in visual math reasoning and document-based question answering. The models are now available to X Premium and Premium+ subscribers through an updated app interface. Unlike some of its competitors in the open weights space, xAI isn't releasing the model weights for download or independent verification. This closed approach stands in stark contrast to recent moves by Meta, which recently released its Llama 3.1 405B model for anyone to download and run locally. xAI plans to release both models through an enterprise API later this month. The company says this API will feature multi-region deployment options and security measures like mandatory multifactor authentication. Details on pricing, usage limits, or data handling policies have not yet been announced. Further Reading Photorealistic image generation aside, perhaps Grok-2's biggest liability is its deep link to X, which gives it a tendency to pull inaccurate information from tweets. It's a bit like if you had a friend who insisted on checking the social media site before answering any of your questions, even when it wasn't particularly relevant. As Mollick pointed out on X, this close link can be annoying: "I only have access to Grok 2 mini right now, and it seems like a solid model, but often seems ill-served by its RAG connection to Twitter," he wrote. "The model is fed results from Twitter that seem irrelevant to the prompt, and then desperately tries to connect them into something coherent."

Grok-2 arrives with image generations -- is the world ready?

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More As anticipated based on updates and new settings in the mobile app for Elon Musk's social network X, a new large language model (LLM) called Grok-2 from Musk's sister company xAI landed last night -- and it's a doozy. Integrated within X itself and available through the Premium ($7 USD/month) and Premium+ ($14/month with no ads) subscription tiers, Grok-2 comes, fittingly, in two model sizes: Grok-2 and Grok-2 mini. Grok-2 offers state-of-the-art performance in a wide range of tasks including chat, coding, reasoning, and vision-based application, while Grok-2 mini is a smaller, faster version optimized for efficiency, suitable for simpler text-based prompts requiring quicker responses. Grok-2 not only boasts image generation capabilities based on a partnership with Black Forest Labs and its new and surprisingly photorealistic open source diffusion AI model Flux.1, but it also shockingly outperforms the AI models from leading rivals including OpenAI (GPT-4o) and Anthropic (Claude 3.5 Sonnet) and even Google (Gemini Pro 1.5) on leading third-party benchmark tests. A new, surprising leader across multiple benchmarks Specifically, Grok-2 and Grok-2 mini outperform all other models on the GPQA, MMLU, MMLU-Pro, MATH, HumanEval, MMMU, MathVista, and DocVQA benchmarks. Even the lmsys-chatbot arena, where many companies covertly test their AI models under alternate names in advance of release (including xAI, where Grok-2 was initially called "sus-column-r") congratulated xAI on the milestone. As AI influencer and University of Pennsylvania Wharton School of Business professor Ethan Mollick observed on X, "There are now five GPT-4 class models: GPT-4o, Claude 3.5, Gemini 1.5, Llama 3.1, and now Grok 2." Musk congratulated his "hardworking xAI team!" on the similarly named social network. Image generations steal the show Even though Grok-2 boasts leading performance on all these different benchmarks related to math, writing, code, and other tasks, by far, the marquee feature capturing the most attention from the jump is its integration with Black Forest Labs' Flux.1 image generation model. Prior to the release of Grok-2, Flux.1 had already been making waves in AI and AI art circles more specifically the last few weeks as people discovered that they could achieve incredibly photorealistic generations from the open source model, enough to resemble familiar situations like a speaker at a TED talk, as well as adapt the model using low-rank adaptation (LoRA) to generate their own likeness in different situations. Now that a version of Flux.1 is integrated directly into Grok-2 much in the same way OpenAI integrated its image generation model DALL-E 3 directly into ChatGPT, allowing users to simply type text prompts to the chatbot and ask it to make them images on command, users are testing this capability out in Grok-2 and finding it is notably permissive -- generating images of even controversial public figures such as U.S. presidential candidates Kamala Harris and Donald Trump, in compromising and fictitious situations. Other leading image generators including Midjourney and DALL-E 3 and Microsoft Designer have prohibitions around generating this type of content -- especially in the wake of the controversy earlier this year over unauthorized explicit deepfakes of popular musician Taylor Swift (made by prompt engineering around the Designer restrictions) -- so it is notable that Grok-2 is bucking that trend and allowing for more freedom, and potential risk. However, that is in keeping with Musk's stated "free speech" ethos for X. Yet users are raising concerns about what the capability means for the providence of deepfakes and misinformation across the web. As user @Omiron33 put it well: "Yes, we've had MJ and Flux, but this is the first to make it usable and quick. Advertising, Propaganda and everything good or bad that comes with that just happened (IMO, the good outweighs the bad)"

Grok-2 is out in beta, now with added AI image generation

The second version of the AI chatbot integrated into X (formerly Twitter), Grok-2 is now being made available in beta to users who pay for X Premium and Premium+. The slowly ingrained AI model, much touted by X owner Elon Musk, comes in two versions: Grok-2 and Grok-2 mini, the latter being described as "our small but capable model that offers a balance between speed and answer quality." In a post on its website, Musk's AI company xAI says Grok-2 has been developed with startup Black Forest Labs to incorporate its FLUX.1 model, which specifically creates AI-generated images. As paying users already with early access to Grok-2 have discovered, the image generator will create images of real people, which raises concerns for artists and copyright, and also doctored imagery -- something Musk's X is actively promoting. Early testers have shared some of their Grok-2 creations on X: Grok can reportedly now generate memes, which have middling results for early testers. xAI also said Grok-2 had been tested on the LMSYS leaderboard -- a system for testing large language models (LLMs) -- and claimed "it is outperforming both Claude 3.5 Sonnet and GPT-4-Turbo" at the time of writing. Grok launched in December 2023 after Musk launched xAI in July 2023. And the AI chatbot hasn't been without controversy. In April, the AI chatbot generated a shocking fake headline reading, "Iran Strikes Tel Aviv with Heavy Missiles." The headline was then promoted on the platform's Explore trending news section.

Meet Black Forest Labs, the startup powering Elon Musk's unhinged AI image generator | TechCrunch

Elon Musk's Grok released a new AI image generation feature on Tuesday night that, just like the AI chatbot, has very few safeguards. That means you can generate fake images of Donald Trump smoking marijuana on the Joe Rogan Show, for example, and upload it straight to the X platform. But a new startup is the one behind the controversial feature. The social media site is already flooded with outrageous images from the new feature. That certainly raises concerns heading into an election cycle, but strictly speaking it's not really Elon Musk's AI company powering the madness. Musk seems to have found a company that sympathizes with his vision for Grok as an "anti-woke chatbot" without the strict guardrails found in OpenAI's Dall-E or Google's Imagen. On Tuesday, xAI announced a collaboration with Black Forest Labs, an AI image and video startup launched on August 1, to power Grok's image generator using its FLUX.1 model. Black Forest Labs is based in Germany, and recently came out of stealth with $31 million in seed funding, led by Andreessen Horowitz, according to a press release. Other notable investors include Y Combinator CEO Garry Tan and former Oculus CEO Brandan Iribe. The startup's co-founders, Robin Rombach, Patrick Esser, and Andreas Blattmann, were formerly researchers that helped create Stability AI's Stable Diffusion models. According to Artificial Analysis, Black Forest Lab's FLUX.1 models surpass Midjourney and OpenAI's AI image generators in terms of quality, at least as ranked by users in their image arena. The startup says it is "making our models available to a wide audience," with open-source AI image generation models on Hugging Face and GitHub. Soon, the company says it plans to create a text-to-video model as well. Black Forest Labs did not immediately respond to TechCrunch's request for comment. In its launch release, the company says it aims to "enhance trust in the safety of these models," however, some might say the flood of its AI generated images on X Wednesday did the opposite. Many images users were able to create using Grok and Black Forest Labs' tool, such as Pikachu holding an assault rifle, were not able to be recreated with Google or OpenAI's image generators. There's certainly no doubt that copyrighted imagery was used for the model's training. This lack of safeguards is likely a major reason Musk chose this collaborator. Musk has made clear that he believes safeguards actually make AI models less safe. "The danger of training AI to be woke - in other words, lie - is deadly," said Musk in a tweet from 2022. Board Director of Black Forest Labs, Anjney Midha, posted on X a series of comparisons between images generated on day one of launch by Google Gemini and Grok's Flux collaboration. The thread highlights Google Gemini's well-documented issues with creating historically accurate images of people, specifically by injecting racial diversity into images inappropriately. "I'm glad @ibab and team took this seriously and made the right choice," said Midha in a tweet, referring to FLUX.1's seeming avoidance of this issue (and mentioning the account of xAI lead researcher Igor Babuschkin). Because of this flub, Google apologized and turned off Gemini's ability to generate images of people in February. As of today, the company still doesn't let Gemini generate images of people. This general lack of safeguards could cause problems for Musk. The X platform drew criticism when AI-generated deepfake explicit images representing Taylor Swift went viral on the platform. Besides that incident, Grok generates hallucinated headlines that appear to users on X almost weekly. Just last week, five secretaries of state urged X to stop spreading misinformation about Kamala Harris on X. Earlier this month, Musk reshared a video that used AI to clone Harris' voice, making it appear as if the Vice President admitted to being a "diversity hire." Musk seems intent on letting misinformation like this pervade the platform. By allowing users to post Grok's AI images, which seem to lack any watermarks, directly on the platform, he's essentially opened a firehose of misinformation pointed at everyone's X newsfeed.

Elon Musk drops Grok 2 -- the X-based AI chatbot is now more powerful and can make images

X has had its own AI chatbot, Grok, for a while, but it'd be fair to say it's not mentioned in the same way that OpenAI's ChatGPT or Google Gemini are. That's not for the want of trying, though, and with a huge user base of X users providing data for the model, a new version was always expected. Now. the obviously-named Grok-2 has entered beta. In a new blog post, X says it represents "a significant step forward from our previous model Grok-1.5, featuring frontier capabilities in chat, coding, and reasoning." "At the same time, we are introducing Grok-2 mini, a small but capable sibling of Grok-2. An early version of Grok-2 has been tested on the LMSYS leaderboard under the name "sus-column-r." At the time of this blog post, it is outperforming both Claude 3.5 Sonnet and GPT-4-Turbo." So, what's new? As the graph above shows, the overall Elo score for an early model of Grok-2 beats out every comparable chatbot except for ChatGPT-4o and Google Gemini. X also says that Grok-2 and its Mini counterpart "achieve performance levels competitive to other frontier models in areas such as graduate-level science knowledge (GPQA), general knowledge (MMLU, MMLU-Pro), and math competition problems (MATH)," while also pointing to vision-based tasks as an area of improvement. Grok will also gain a new interface on X, as well as the option to generate images with a prompt. This is achieved through the integration of the popular Flux AI image generation model from Black Forest Labs. Grok will be offered through a new enterprise API later this month, which X is promising will offer a "bespoke tech stack" as well as mandatory multi-factor authentication.

Elon Musk's xAI unveils new Grok-2 and Grok-2 mini models

Generative artificial intelligence startup xAI unveiled the newest versions of its Grok models, Grok-2 and Grok-2 mini, as the competition between AI companies rages on. "We are excited to release an early preview of Grok-2, a significant step forward from our previous model Grok-1.5, featuring frontier capabilities in chat, coding, and reasoning," xAI said in a blog post. "At the same time, we are introducing Grok-2 mini, a small but capable sibling of Grok-2. An early version of Grok-2 has been tested on the LMSYS leaderboard under the name 'sus-column-r.' At the time of this blog post, it is outperforming both Claude 3.5 Sonnet and GPT-4-Turbo." Claude 3.5 Sonnet is one of Anthropic's models, while GPT-4 Turbo was created by OpenAI. Google's (GOOG) (GOOGL) Gemini generative artificial intelligence chatbot recently said OpenAI's GPT-4 is the most advanced large language model, according to investment firm Baird. Both Grok-2 and Grok-2 mini are available in beta testing on the X (formerly known as Twitter) social network, xAI added. The two models will also be available via the enterprise API later this month. Earlier this week, OpenAI, which is backed by Microsoft (MSFT), unveiled a new version of its GPT-4o large language model, but provided little detail on it. A number of investors have put money into xAI, which hit a $24B valuation in May after it raised $6B in a new fundraising round. Investors in that round included venture capital firms such as Valor Equity Partners, Vy Capital, Andreessen Horowitz and Sequoia Capital. Other investors include Fidelity Management & Research Company, Prince Alwaleed Bin Talal and Kingdom Holding, amongst others. Musk heads up xAI, along with X, Tesla and several other companies.

Elon Musk's Grok-2 Beta Launched; Outperforms ChatGPT, Claude, and Gemini

Users can start using the new model on x.com, but it will require an X Premium subscription. Elon Musk's AI venture, xAI has released an early preview of the Grok 2 model, and it has surprisingly outperformed Claude, Gemini, and even ChatGPT as well. The earlier Grok-1.5 model was not received well, but Grok-2 has delivered great performance on the LMSYS leaderboard. xAI has released two new models: Grok-2 and a smaller Grok-2 mini model. xAI says Grok-2 has been significantly improved in key areas including reasoning, instruction following, and providing accurate and factual information. In traditional AI benchmarks, Grok-2 has scored a whopping 87.5% in MMLU and 88.4% in HumanEval. This is particularly interesting because the MMLU score has been derived using 0-shot CoT. Grok-2 was tested on LMSYS under the name "sus-column-r". With around 12,000 votes, it stands at the third position, just below ChatGPT-4o-latest, Gemini-1.5-Pro-Experimental, and GPT-40-2024-05-13. However, it performs better than GPT-4o-mini, Claude 3.5 Sonnet, Gemini 1.5 Pro, and Llama 3.1 405B. In coding and math-related tasks, Grok-2 takes the 2nd spot, and in hard prompts, it takes the 4th position. xAI says that the Grok-2 multimodal model will be released soon. The company has not revealed the parameter size for both models. You can start using the new Grok-2 model on x.com and developers can get started with the API as well.

xAI Grok-2 and Grok-2 mini beta out on X

Elon Musk's xAI announced the early preview of Grok-2 and Grok-2 mini, an upgraded version of its existing Grok-1.5 model. These new AI assistants come with advanced capabilities in both text and vision understanding, integrating real-time information from the X platform, accessible through the Grok tab in the X app. The company is said to have evaluated the models in two key areas: following instructions and providing accurate, factual information. Grok-2 has shown significant improvements in reasoning with retrieved content and in its tool use capabilities, such as correctly identifying missing information, reasoning through sequences of events, and discarding irrelevant posts, said the company. Grok-2, under the name "sus-column-r," is now outperforming both Claude 3.5 Sonnet and GPT-4-Turbo at the LMSYS chatbot arena, a popular competitive language model benchmark. The new model excels in reasoning, reading comprehension, math, science, and coding, showing significant improvements over Grok-1.5. With a redesigned interface and new features, X users can use Grok-2 and Grok-2 mini for seeking answers, understanding the context of memes, enhancing search capabilities, gaining deeper insights on X posts, and improving reply functions, as well as collaborating on writing or solving coding tasks. The company is also experimenting with Black Forest Labs FLUX.1 model to expand Grok's capabilities. Grok-2 and Grok-2 mini are currently available in beta on X, and users are urged to update their X app to the latest version to experience it. The company is also making both models available through our enterprise API later this month.