1. AI image generator Midjourney has stopped offering free trials due to an influx of users creating throwaway accounts to access the service.
2. The company's CEO denies that the pause is linked to recent viral images created using Midjourney, including fake pictures of Donald Trump being arrested and the pope wearing a stylish jacket.
3. Midjourney's content restrictions are more permissive than some rival services but more restrictive than others, and the company is working on improving its moderation policies.
The article discusses the decision of AI image generator Midjourney to halt free trials of its service due to an influx of new users. The author provides insights into potential biases and sources of one-sided reporting, such as the initial assumption that the pause was linked to a spate of recent viral images created using Midjourney, including fabricated images of Donald Trump being arrested and the pope wearing a stylish jacket. However, the CEO clarified that the free trial never included access to the latest version of Midjourney, version 5, which creates the most realistic images and was used for these viral pictures.
The article also highlights Midjourney's response to politically sensitive escalating threats from fake images, which has been piecemeal without any significant overhaul of its moderation policies. The author notes that Midjourney's content restrictions are more permissive than some rival services but more restrictive than others. However, there is no exploration of counterarguments or evidence for claims made in this regard.
The article also mentions Midjourney's list of banned words related to topics in different countries based on complaints from users in those countries. Still, it doesn't share a complete version of this list to minimize drama. The author notes that such restrictions are easy to bypass but doesn't explore possible risks associated with this.
Overall, while the article provides valuable insights into Midjourney's decision to halt free trials and its response to fake images' threats, it lacks exploration of counterarguments and evidence for claims made. Additionally, it could benefit from presenting both sides equally and noting possible risks associated with content restrictions' ease of bypassing.