DALL-E AI Art Generator Finally Opens the Doors to a Wider Internet

DALL-E AI Art Generator Finally Opens the Doors to a Wider Internet

AI-generated artwork showing a man holding his chest while typing on a computer

I asked DALL-E to create an image of a journalist struggling to meet deadlines in the style of Caravaggio. I wouldn’t say it really captured the old master’s flair, it’sit’s far too impressionistic. However, I feel the pain of the subject.

Art and image archives on the Internet are already flooded with images developed using artificial intelligence. Expect even more high-imaginative images or photos of dubious origin now that the AI ​​image generator that arguably started the current artificial imagery craze, DALL-E, is open and accessible to everyone. all.

In a Wednesday blog postthe DALL-E OpenAI developer said so has already 1.5 million users create more than 2 million AI-generated images per day. Using data and feedback, the company said it has tightened its filters to reject any images designed to mimic sexual, violent or polthisical content. There is no API currently available for DALL-E, but apparently one is in development.

The DALL-E section of OpenAI website always asks users to join a waiting list, from the moment of reports, although there is now a sign to the top. In an emailed statement, OpenAI said it took “an iterative deployment approach from the start to evolve DALL-E responsibly, which helped us discover how it can be used as a powerful creative tool.” .

Users who sign up receive 50 free credits to create images in the first month, then 15 free credits each month thereafter.

OpenAI’s image generator was first revealed in April and people quickly joined the waiting list, with some twiddling their thumbs for months before getting their turn. Although DALL-E – named after famous artist Salvador Dali and written as Disney Pixar WALL-E– was the first system to take a real leap forward in the capabilities of AI image technology, other systems have caught up, at least in terms of popularity. Mid-term hosts hundreds of thousands of users on its Discord-based platform, and StabilityAI, the creators of the AI ​​art generator Stable Diffusion, discussed raise millions on the back of his looser and more controversial system.

The OpenAI announcement comes in a very strange place due to both the growing popularity of AI arts, but also the public backlash against it. The Washington Post spoke with several OpenAI product managers while showing how the software could be used to create images of fake protests, which would go against the company’s restrictions on creating political images. The system limits user prompts by triggering content warnings on words such as “tween” and “teen”. At the same time, even though the system is expected to restrict prompts based on public figures, the Post noted that it still allows users to generate images based on people like Mark Zuckerberg and Elon Musk.

And there is always a major question of ownership. A technical manager has become a figure of consternation for entering an AI-generated artwork in a local art contest and win first prize. Last week, an artist claims she received the first copyright for a work created using AI, but the US Copyright Office said it does not accept any work that was not created by human hands, which which means the matter remains in limbo.

Of course, none of the most popular image generators have avoided controversy. Stable Diffusion has been cited to be used for generating child pornography, but StabilityAI founder Emad Mostaque said they were working on systems to block such content. Business leaders at StabilityAI and OpenAI have even gone back and forth over which of their systems is less controversial.

Last week OpenAI announced that they were lifting of restrictions which prevented users from uploading real human faces for the AI ​​model to start editing. The company promised to create detection technology to prevent users from abuse the system to create porn or violent content. Users were reportedly prevented from uploading photos of people’s faces without the subject’s consent. The company had previously opened its systems researchers seeking to create artificial human faces.

#DALLE #Art #Generator #Finally #Opens #Doors #Wider #Internet

Leave a Comment

Your email address will not be published. Required fields are marked *