Midjourney, a generative artificial intelligence platform that allowed users to generate images from text commands, will no longer be free due to the large number of people creating throwaway accounts to get free images. According to The Washington Post, the company's CEO, David Holz, has announced on Discord that free trials have been "disabled" due to "extraordinary demand and abuse of judgment" from users. Users who want to generate images using AI with this platform will have to pay between $10 and $60 per month for the service.
The danger of new generative artificial intelligence tools is obvious. Midjourney has behind it a generative AI that does so well that the images look real, which can contribute to misinformation and fake news on the Internet. The company's own CEO has indicated that in the future there will be moderation tools that will be able to find generated images that were created in violation of usage policies, although it is clear that at the moment the rules and moderation are not working at the highest level of refinement.
Midjourney became popular after the AI exploded on the Internet, flooding everything with its software to generate content. Since then, we have seen the creativity of Internet users grow, but there have also been cases of inappropriate use of these tools. In this case, the Midjourney platform has had to stop being free due to the abuse of some users who create disposable accounts to obtain free images.
The Midjourney platform has earned the reputation of being one of the best platforms for create images from text commands thanks to AI. However, the realism of the images generated by this platform is such that there is an obvious danger that society will take the images generated as truthful. That is why the company is having trouble determining content policies on generating images of real people. The CEO of the company pointed out that in the future there will be moderation tools that will be able to find generated images that were created in violation of the usage policies, although at the moment the rules and moderation are not working at the highest level of refinement.
Currently, text-based generative AIs - such as Google's Bard or ChatGPT - are more stringent than image-based AIs, such as Midjourney or Stable Diffusion. This is because text-based generative AIs are trained on much larger and more diverse datasets, while image-based generative AIs still struggle to generate high-quality, realistic images.
Leave a Reply