One of the issues we talked about most in the past months was Dall-e, an artificial intelligence technology that instantly transforms texts into images. Developed by a company called OpenAI and announced in 2021, artificial intelligence was recently put into service by around 1 million users. Now, there has been an important development regarding Dall-e.
OpenAI blocked the use of faces in Dall-e in the first stage. The reason was to prevent malicious users from using real human faces. However, this decision has now been abandoned. In the statement made by OpenAI, it was stated that Dall-e 2, which is a more advanced structure than the original Dall-e, is allowed to use and edit faces. Thus, users will be able to use human faces in images created from text.
Abuse is said to be minimised
Sending an e-mail to its users on the subject, OpenAI says that they have received a lot of feedback for face editing and that they are working on this issue and developing new detection and response techniques. These new techniques have minimized the abuse of the face use feature. Also updating the Dall-e 2 terms of use, the company says that users will not be able to upload face photos to the system that they do not have the rights to. We will see how useful this will be in practice in the future.
What would happen if faces were used in the texts depicted?
To be frank, OpenAI’s previous decision was more correct. Because we have seen in deepfake technology how dangerous it is for artificial intelligence to use real people’s faces. There will be those among you who remember; With deepfake technology, fake content belonging to the world’s most famous names was created and these contents were served over the internet. For now, it is unknown what kind of results Dall-e 2’s face use feature will cause…
16488
