The block comes after reports of users on platforms like 4chan using free tools like Bing Chat and Stable Diffusion to create and distribute racist images. Microsoft did not provide details on what the new guardrails or filters could look like and did not comment on whether it recently blocked content related to the Twin Towers. The developers of DALL-E 3 have acknowledged that its safety measures are not perfect and are constantly being upgraded.
Key takeaways:
- Users of Microsoft's Bing Chat and Bing image generator, integrated with OpenAI's DALL-E 3, found a loophole to create images of animated characters and the Twin Towers, leading Microsoft to block prompts related to the Twin Towers.
- AI image generators have been under scrutiny for copyright claims and deepfakes, leading developers to be more cautious about the content that can be created.
- Microsoft plans to improve its systems to prevent the creation of harmful content, implementing guardrails and filters for Bing Image Creator, though the specifics of these measures have not been detailed.
- Despite these measures, users have been guided on how to manipulate free tools like Bing Chat and Stable Diffusion to create and distribute inappropriate images, highlighting the ongoing challenges in regulating AI-generated content.