In a nutshell, it gives AI the ability to study what individuals or objects look like, then synthesizes a “photorealistic” image of the subject in a completely new context.However, as multiple experts told me, DreamBooth may soon facilitate the creation of deepfakes indistinguishable from real photos or videos, at scales never before seen.BeFake uses DreamBooth in conjunction with another popular AI image generator called Stable Diffusion, Wahlforss said: “To get it perfectly right, you have to add algorithms on top or fine tune it in a certain way.” To create my images, I sent Wahlforss 20 close-up photos of myself in varying backgrounds and clothing.“On our end, we have a lot of checks and balances, and that’s one of the algorithms that I was talking about … obviously we’re very cautious around generating anything that is not safe for work, either violence or nudity.” “But I don’t want to go into too much detail because then, obviously, you’re opening yourself up to attacks.” The team behind DreamBooth did not respond to the Star’s requests for comment before publication.AI, which created AI image generator Stable Diffusion, DreamBooth is “just the beginning of a new wave of generative models in image, language, video, audio and 3D.” The technology is not limited to 2D images, he said — image generation models are already being used in animation and “techniques for fine-tuning such as Dreambooth will absolutely be possible with other modalities, further extending the options.” In response to the many people making use of the open source technology, Mason said: “We encourage ethical use of the models and a sensible open discussion with transparency on the techniques and applications that the community is developing.” The future of deepfakes?"