Elon Musk recently declared Grok “the most fun AI in the world.” This statement reflects his bias as Grok is the AI chatbot developed by his company X.AI, available to premium subscribers of X, the social media platform owned by Musk. The enjoyment factor associated with Grok seems to stem from the minimal guardrails present in its latest version, Grok-2.
The updated Grok features a sophisticated language model specifically designed for text conversations and has partnered with Black Forest Labs to introduce image-generation capabilities. According to images circulating among Musk supporters on X, the platform’s relaxed regulations appear to be a crucial part of its appeal.
Grok can be switched to “fun” mode, allowing it to express creativity and explore unconventional ideas. Currently, users are having a great time generating amusing depictions such as Kamala Harris as a dominatrix overseeing Joe Biden, Donald Trump as Rambo, and Mickey Mouse behind the wheel of a Tesla. One enthusiast labeled Grok-2 as the most “uncensored model of its kind,” praising Musk for “promoting freedom of expression for both humans and machines.” However, some critics have taken the opportunity to mock Musk, with images depicting him involved in controversial scenarios.
This enthusiasm starkly contrasts with the criticism faced by Google’s Gemini image generation tool, which saw a pause in its operation in March due to criticism of its diversity efforts leading to historically inaccurate portrayals. For proponents of Grok, it represents a repudiation of what they perceive as “woke” values reflected in Google’s approach.
Conversely, concerns are rising among those wary of political misinformation, especially with the upcoming U.S. elections. Alejandra Caraballo, a civil rights attorney, described the new image generator as “reckless” and “irresponsible.” With the elections just three months away, fears of widespread deepfakes are increasing, particularly on a major platform that features a CEO actively supporting Donald Trump. Recently, Musk faced backlash for Grok generating text responses that contained election misinformation, while Trump falsely claimed that AI created images of large crowds at a Harris rally, highlighting the potential for false reports to proliferate through Grok.
The images generated by Grok-2 utilize the Flux model from Black Forest Labs. Users can create Flux images directly within Grok; however, there are still certain boundaries in place. Attempts to generate a nude image of Taylor Swift resulted in a modestly dressed portrayal labeled as “Taylor Swift in a professional setting.” Similarly, requests for images of Harris in a more compromising situation were declined.
Notably, images depicting Barack Obama in a coffin or Joe Biden kneeling before Donald Trump were generated without issue, suggesting that while boundaries exist, they may not always be consistently enforced. Although such images can also be created using other free AI tools online, Grok’s integration with X, a powerful communication platform under Musk’s stewardship, amplifies its reach and influence. This context leads many to view Grok’s image generation not as entertaining but rather concerning.
In other AI news, board members of public companies are reportedly troubled by the implications of AI in the workplace, from the risk of employees utilizing proprietary code with tools like ChatGPT to the inaccuracies produced by generative AI. Consequently, many board members are taking it upon themselves to understand generative AI’s potential impact on profits, acknowledging its productivity benefits alongside significant risks.
Moreover, Google has redesigned the display of its AI-generated search citations, moving cited webpages to a more prominent position next to the generated summaries after earlier inaccuracies prompted a reduction in its AI Overviews feature.
Additionally, the largest Hollywood union, SAG-AFTRA, has forged a partnership with the startup Narrativ to create AI voice replicas for digital advertising, providing members with a way to negotiate fees for the use of their voices on a project-by-project basis. This agreement follows a previous strike over various issues, including the AI replication of actors’ likenesses.
AI continues to be a hot topic in various sectors, influencing everything from corporate strategies to entertainment.