MidJourney Halts Free Trials Amidst Deepfake Misuse
MidJourney, an AI image generator, has decided to end free trials of its service due to increasing abuse by users. The company’s CEO, David Holz, explained that the demand and trial abuse were extraordinary, and current safeguards were not sufficient to prevent misuse during trial periods.
High-Profile Deepfakes Lead to Concerns
The AI image generator has attracted unwanted attention after users created deepfakes of Donald Trump being arrested and Pope Francis wearing a trendy coat. While these images were quickly identified as fake, there is growing concern about potential misinformation caused by similar generators like MidJourney and OpenAI’s DALL-E.
Struggling to Establish Content Policies
MidJourney has found it difficult to set content policies as their AI enables more realistic imagery. In an effort to minimize drama, the company banned images of Chinese leader Xi Jinping in 2022. The founder revealed plans to improve AI moderation to screen for abuse.
Developers Adopt Strict Rules to Prevent Misuse
Some developers, like OpenAI, have imposed strict rules to avoid misuse of their technology, barring images of ongoing political events, conspiracy theories, and politicians, as well as hate, sexuality, and violence. Others, like Stability AI, have relatively looser guidelines for their users.
AI Image Production Faces Additional Challenges
Misleading content isn’t the only issue with AI image production; there are also concerns about image theft as AI-generated art often uses existing images as reference points. This has led to hesitancy from companies worried about unwanted attention from using AI art in their products.