**Black Forest Labs: Powering Grok’s Image Generator**
German startup Black Forest Labs recently partnered with Elon Musk’s Grok to power its AI image generator using the FLUX.1 model. Black Forest Labs, which came out of stealth with $31 million in seed funding, aims to support Musk’s vision for Grok as an “anti-woke chatbot” without the strict safeguards found in other AI image generators. The collaboration between the two companies has already resulted in a flood of outrageous and controversial images being uploaded to the platform.
**Superior Quality Models**
Artificial Analysis ranks Black Forest Lab’s FLUX.1 models as superior in quality compared to Midjourney’s and OpenAI’s AI image generators. Users on the platform have rated Black Forest Lab’s models highly in terms of image generation quality. The startup plans to make their models available to a wider audience through open-source AI image-generation models on platforms like Hugging Face and GitHub. Additionally, they have expressed intentions to develop a text-to-video model in the near future.
**Lack of Filters and Controversial Images**
Grok’s image generation feature lacks filters, leading to the creation and upload of controversial and potentially offensive images. Users have been able to generate images that were not possible with other image generators, such as Pikachu holding an assault rifle. It is apparent that copyrighted imagery was used in training the model. While the startup claims to aim for enhanced trust in the safety of their models, the flood of AI-generated images has raised concerns about the lack of safeguards in place.
**Musk’s Stance on Safeguards**
Elon Musk’s choice to collaborate with Black Forest Labs and their lack of safeguards align with his belief that AI models should not be “woke” or restricted by safeguards. Musk has emphasized that training AI to lie is dangerous. This collaboration allows Grok to push boundaries and generate images without strict limitations, giving users more freedom in what they can create.
**Comparison with Google Gemini**
Black Forest Labs’ FLUX.1 model has been praised for avoiding the issues faced by Google Gemini, which had difficulty generating historically accurate images of people. Gemini faced criticism for injecting inappropriate racial diversity into images. Black Forest Labs’ board director, Anjney Midha, highlighted this difference and commended the team for their approach. Google took measures to address the issue and disabled Gemini’s ability to generate images of people. In contrast, Grok’s collaboration with Black Forest Labs appears to have sidestepped these problems.
**Misinformation and Controversy**
Grok’s lack of safeguards has resulted in the spread of misinformation and controversy on the X platform. Deepfake explicit images of Taylor Swift and AI-generated headlines have garnered attention and criticism. Secretaries of state recently urged X to stop spreading misinformation about Kamala Harris. Musk himself has shared a video that used AI to clone Harris’ voice, contributing to the propagation of false information. By allowing users to post Grok’s AI images without watermarks directly on the platform, Musk has inadvertently enabled a flood of misinformation on users’ newsfeeds.