Following my recent success in using the AI image-generation program Adobe Firefly (AF) to generate video clips after receiving from me both verbal and visual prompts (click here to view a selection of these clips), I decided to subject its ability to further, more eextreme scrutiny, in order to test just how creative it could be, if given the opportunity. I chose to do this by providing it with a single still picture and, as the ultimate test of AI 'imagination', just a single word as its verbal prompt from me.
Knowing that dancing, sadly, has never been one of my talents, I decided to turn this deficit to my advantage now, by making it the subject of this particular AI test. So, what I did was provide AF with a photo-still image from my real-life (not AI-generated) Wild Youth video clip showing me wearing my biker gear and smoking a ciggie back in my late 20s that I recently posted here, together with the single word 'jiving', to see what, if anything, it could generate from such a paucity of prompts, with me having pared them down to the absolute minimum of just one picture and just one word.
To say that I was startled by the resuls would be putting it mildly! As seen from the video clips that AF generated from those prompts and which are now presented here, not only am I animated as a mover and groover far beyond anything that I have ever accomplished in reality, but, without any prompting whatsoever from me, it also created a number of fully-formed, fully-functional dancing buddies jiving and gyrating alongside me! But from where did AF derive their descriptions? (And why were no lady dancers included among them?) I can only assume that these figures constitute some kind of default characters, programmed into its database and automatically generated in the absence of specific descriptive details having been provided to it via verbal or visual prompts.
As you'll see, the vid clips are totally hilarious - especially the final one in which after momentarily turning my back to the camera I spin round again, only to reveal that I have transformed into a completely different person! - and outrageously camp (what that says about the AI program's 'personality' is anyone's guess!). But they are also thoroughly fascinating, and not a little perplexing, even somewhat disturbing, too. For, suddenly, I have seen what an AI image-generation program can achieve if left almost entirely to its own devices, and what it can achieve is well beyond anything that I'd anticipated. Remember: apart from me, the curtains & bed, everything & everyone contained in these vid clips were created entirely by the AI program - nothing & no-one originated from me at all. So, should we be worried about the capabilities of ever-advancing AI? Or, to rephrase this in a somewhat more apt manner, are we dancing ever more closely towards the date of our own demise as our planet's dominant species? Sometimes silly vids can give rise to serious questions, and this may well be one such time...
And here are all of the above individual vid clips linked together to yield a single continuous video:
More of my AI-created biker images next time!
Plus: please click here to access a chronological listing of fully clickable links to all galleries uploaded by me on this blog. NB - all of the video clips included in this page's gallery were created by me, RebelBikerDude, using the AI image-generation program Adobe Firefly, for non-commercial, entertainment purposes only, and do not feature any real-life persons or copyrighted characters.
No comments:
Post a Comment