Google's Relightables system makes it possible to customize lighting on characters in real time or re-light them in any given scene or environment. (Project Page)
Technically this has been available in some form since the mid-80s and is of very limited practical use, but … all on board on the "creative AI" bullshit hype train!
“To the extent that propaganda is based on current news, it cannot permit time for thought or reflection. A man caught up in the news must remain on the surface of the events; he is carried along in the current, and can at no time take a respite to judge and appreciate; he can never stop to reflect ... One thought drives away another; old facts are chased by new ones. Under these conditions there can be no thought. And, in fact modern man does not think about current problems; he feels them. He reacts, but he does not understand them any more than he takes responsibility for them. He is even less capable of spotting any inconsistency between successive facts; man's capacity to forget is unlimited ... This situation makes the 'current-events man' a ready target for propaganda.” — Jacques Ellul, Propaganda (1973)
In his books Empire and Communication (1950) and The Bias of Communication (1951), the Canadian historian Harold Innis argued that all communication systems incorporate biases, which shape how people communicate and hence how they think. These biases can, in the long run, exert a profound influence over the organization of society and the course of history.
Is there an overarching bias to the advance of communication systems? Technology enthusiasts like Kelly would argue that there is — a bias toward greater freedom, democracy, and social harmony. As a society, we’ve largely embraced this sunny view. Harold Innis had a very different take. “Improvements in communication,” he wrote in The Bias of Communication, “make for increased difficulties of understanding.” He continued: “The large-scale mechanization of knowledge is characterized by imperfect competition and the active creation of monopolies in language which prevent understanding and hasten appeals to force.” Looking over recent events, I sense that Innis may turn out to be the more reliable prophet.
If you find current generative media developments confusing (deepfakes, HQ generative music, etc.), you'll be truly confused by the mid 2020s when computational & biological generative systems start to heavily intertwingle. Better start to cultivate a meditation practice today.
A while back, got to contribute a tiny bit to an upcoming book by MIT researchers around the topic of Co-Creating media with groups of people and algorithms. Just got to check the draft and happy it turned out really fascinating. Should be out in 2019.
