At the beginning of the year we discussed the powerful qualities of GIVE HER, an artificial intelligence system that through a neural network can generate images based on a textual description.
That tool is not available ready to be used from the web, but the technology behind it is freely accessible. A web portal, using a similar AI mechanism, posed an interesting interactive dynamic to take advantage of this tool.
Art by An AI, a collaborative digital art gallery
Art by the AI it is presented as a daily competition. Every 24 hours, the ten images displayed on its cover are renewed, all generated by Alice, its AI.
Using the VQGAN and CLIP algorithms, Alice is able to generate images based on a text description, under a dynamic similar to that of DALL-E. The interesting thing is that the choice of phrases or concepts to be used is defined based on an open vote. Any visitor can submit their proposal, hoping to obtain enough votes during the day to appear among the top ten preferences.
At 12:00 UTC each day, the images are removed from the platform forever, in order to give them a certain exclusivity. Hand in hand with this aspect, the published illustrations are shared compressed, in a reduced resolution, since only during their exhibition is offered the possibility of acquiring a poster or canvas with a high quality printed duplicate, enlarged by AI, provided that the result generated is of acceptable quality.
Within the versatility with which it has been shown that it is possible to use artificial intelligence, the generation of images through artificial intelligence has been one of its most striking aspects, but at the same time difficult to access in its maximum splendor from the sidewalk of end users. Just as DALL-E was presented primarily at the conceptual level, for use in research and development environments, other tools such as This Person Does Not Exist They allow their users to request the generation of new images (of faces, in this case), but without the possibility of altering any parameter.
In the case of this virtual art gallery, although participation is not totally direct and free, a participatory component is added that allows, at least, to try your luck participating. In addition, if the idea proposed is popular enough, an instance is opened to verify if the idea initially conceived in our mind is close to what this AI is capable of generating.