Tools for Humanity, which was co-founded by OpenAI’s Sam Altman, has launched its World ID system, aimed at differentiating real humans from AI chatbots. The process involves users signing up for an in-person iris scan using Worldcoin's orb, a device the size of a bowling ball. Once the scan confirms that the individual is human, a World ID is created.
This is a part of Tools for Humanity’s larger Worldcoin project. After securing two million users during its beta period, Worldcoin is now expanding its scanning operations to 35 cities across 20 countries. The World ID is stored on blockchain networks, which guarantees privacy and prevents control by any single entity, according to co-founder Alex Blania.
The World ID system can help distinguish real people from advanced AI chatbots like ChatGPT. The system has received endorsement from Binance, the largest cryptocurrency exchange, which plans to list Worldcoin for trading.
In addition, co-founder Altman believes that Worldcoin could contribute to tackling economic issues brought by the increasing prevalence of AI. Specifically, the World ID system could be a potential tool for implementing universal basic income (UBI) and reducing fraud in its distribution, given that only real humans can have World IDs.
Although Altman acknowledges that such a world may be quite distant, he sees Worldcoin as laying the necessary groundwork for its realisation.
Stable Doodle lets you turn your sketches into art for free
Stability AI, the company behind the image-generating tool called Stable Diffusion, is launching a new service that turns simple sketches into full images. This service, named Stable Doodle, creates fully rendered images using a simple drawing and a text description of what you want the final image to look like. It is currently available through ClipDrop, a platform that Stability AI recently acquired.
Stable Doodle is user-friendly and can be used by both professionals and beginners, even those who are not tech-savvy. All you need is the ability to draw basic sketches and internet access, and you can create high-quality images in just seconds.
The ability to guide a final product using a basic sketch or outline is a result of the way diffuse models work. Diffusion models, the type of AI behind modern image generators like Stable Diffusion and Midjourney, create images starting with random noise. Over a number of steps, the diffusion model creates images that look more and more like the text prompt it was given.
Giving the model something other than random noise to start with can be very effective at guiding the process in the direction you want. The prompt “kick” could look like almost anything. But by giving the model an outline, perhaps in the form of a stick figure and a circle, the model is far more likely to create an image that looks like a person kicking a ball.
Stability AI hopes that Stable Doodle will help professionals such as designers and illustrators save time and work more efficiently. However, they also note that the quality of the final image largely depends on the detail of the original sketch, the descriptiveness of the prompt, and the complexity of the scene being depicted.
Stable Doodle is free, but there are limits to its use. The launch of Stable Doodle is part of Stability AI's efforts to generate more income after a slow period in its business operations. Despite this, Stability AI has recently raised $25 million, bringing its total funding to over $125 million. The company's last valuation was at $1 billion, and reports suggest that they are aiming to increase this valuation fourfold in the coming months.
James Browning is a freelance tech writer and local music journalist.