© IE Online Media Services Pvt Ltd
Tags:
Journalism of Courage
One of the prominent capabilities of a generative AI model is the ability to morph any picture. However, most AI tools have certain restrictions on what they can edit/replace in any image, and a new tool by Hugging Face takes the AI editor to the next level by enabling users to replace or generate any part of the image with just a simple text prompt.
Hugging Face’s ReplaceAnything AI image editor is currently available for the public for free, which allows users to upload any picture and add or remove a specific part of the image using a simple text prompt. Users can either input their own prompt or select one from the pre-defined ones to get more accurate output.
One can upload an image of a user holding a specific product and can ask AI to either remove or replace that product using a text prompt. ReplaceAnything can replace face, objects, background, skin tone, and can also remove the background of an image.
Images generated using ReplaceAnything are known as AIGC (AI-generated content), and as generative AI progresses, products like these could get better, which will lead to an increase in deep fake content, which will be indistinguishable from the original image.
Deep fake content is already on the rise, where, recently deep fake images of Tailor Swift (edited using AI tools) went viral, and we have also seen similar developments in India, which according to 404 Media, were linked to a specific Telegram group, where, Microsoft’s text-to-image generator were one of the tools used by the group members.
While most generative AI tools have certain regulations on what commands they can respond, however, that isn’t fully stopping people from using these tools to create morphed pictures of videos.
