[ad_1]
Be part of us in Atlanta on April tenth and discover the panorama of safety workforce. We’ll discover the imaginative and prescient, advantages, and use instances of AI for safety groups. Request an invitation here.
ChatGPT’s picture technology capabilities simply acquired an improve. OpenAI just posted on its account on the social network X an replace declaring that customers of GPT-4 powered ChatGPT can now click on on photos generated inside the chatbot and use a paintbrush fashion instrument to spotlight sections of the picture they need to regulate or add to.
In a video displaying off the capabilities, a person generates photos of a canine with a conical social gathering hat atop it.
The person can then be seen clicking the picture to broaden it and enter a brand new editor view, the place they choose a paintbrush icon on the prime, brush over parts of the canine’s head, after which sort in a brand new immediate “add bows,” which preserves the unique picture however generates a brand new model with crimson bows showing within the user-highlighted area.
A number of internet customers famous that they had noticed the modifications pushed stay quietly earlier this week and final.
VentureBeat tried out the brand new function on a private account and it labored as described.
Powered by DALL-E 3 and one other new replace
OpenAI’s picture technology function inside ChatGPT is powered by its underlying DALL-E 3 mannequin which was unveiled by the company back in September 2023. The addition of DALL-E 3 instantly inside ChatGPT drastically expanded the previous mannequin’s capabilities for customers, and the brand new addition at present once more additional extends its utility and fine-grained management for customers.
VentureBeat customers ChatGPT/DALL-E 3 and different AI artwork turbines to create article header photos and different paintings for our properties.
OpenAI additionally tweeted that it has added a brand new fashion inspiration function in its DALL-E 3 GPT, permitting customers to pick out between a number of pre-set kinds akin to woodcut, close-up, low angle, synthetic lighting and hand-drawn.
The strikes come even amid ongoing legal challenges to OpenAI’s models and rising pushback by some artists and web users against AI art generally, particularly fashions like DALL-E 3 that had been presumably educated on copyrighted imagery with out specific permission, consent or compensation.
[ad_2]
Source link
