Bob Ross lives
|Bob Ross lives|
|Name||Bob Ross lives|
By Lenka Hamosova & Pavol Rusnak
Bob Ross Lives!
Creative potential of GANs
In the context of the latest development in deep learning and specifically in Generative Adversarial Networks (GANs), we read a lot about the negative side of these tools being widely accessible. The media focus on harmful deep fakes and their consequences on politics or privacy. The creative potential of visual generative power of the AI is left behind.
Let’s embrace the positive side of the neural networks' ability to generate synthetic imagery and explore the potential of these tools with a diverse group of curious makers. Before large companies take over and colonize this creative space, or it gets distilled into simple entertaining Snapchat filters, we want to invite hackers and designers to get their hands dirty and experiment with the deep learning models that would have been Bob Ross' wet dream!
The future brings a massive challenge for fast adaptation of our senses to the new visual reality as well as the necessary adaptation of our work methodologies. Formerly impossible is now possible - artificial intelligence can generate photorealistic sceneries, objects, animals and humans that are not part of this physical world. Despite its worrying nature, that it might lead to the state of confusion in communication (distinguishing between what’s real and what’s fake), this technological progress also means unforeseeable advancements in the work of visual makers. It’s a new territory that deserves exploration right now, while it’s still evolving. Especially for creative professionals, this evolution means necessary reimagining of their practice, discovering new methodologies and appropriating unexpected new tools. Video might have killed the radio star - GANs might as well kill a lot of creative visual jobs, but will definitely liberate many from laborious visualising in favor for more conceptual and valuable work!
10:00 introduction to GANs and currently available creative AI tools
11:00 “You need the dark in order to show the light.” start of the workshop - part 1 (speculative phase)
12:00 lunch break
12:30 “Let’s get a little crazy here!“ workshop part 2 (production phase)
15:00 Congratulations! You’ve just become Bob Ross of synthetic visual media! (presentation of the results & discussion)
16:00 end of the workshop
- Note that this workshop is happening in parallel to another one, you will have to sign up!
Workshop PART 1: speculative phase
“You need the dark in order to show the light.”
In the speculative part of the workshop we will discuss all kinds of future scenarios for the use of synthetic visual media, their ethical consequences as well as creative potential and set the base for productive exploration in the 2nd part of the workshop.
1. Everyone choose the side! Are you a white knight or is the dark side calling you? Are you interested more in the potential conflicting moments of fakes, or do you prefer to focus on how generated visuals could change your creative methodology?
- we split in two groups
2. Brainstorm your chosen scenario with the group. Try to define the use case for your future synthetic visuals. Search for conflicting situations - it does not have to have extreme consequences, can also be personal!
3. Think of a parallel in the history of visual communication, when it was not possible to be sure which visual information were true. How did we validated visual information back then? Think of an authenticating strategy for your specific scenario.
Workshop PART 2: production phase “Let’s get a little crazy here!”
In the production phase we will get our hands dirty and our virtual brushes beaten. Experiment with various available generative neural networks tools and stretch the limits of options they offer. If you believe in it, you can do it, Bob Ross used to say. Produce the desired generated visual for your chosen scenario, if possible.
1. Play with available models and try to generate fake visuals for your chosen scenario
2. Experiment with changing the values and inputs. Random or designed, accidental or conceptual - all methods count! Document the process, including the fails.
- Remember: “We don't make mistakes, just happy little accidents.”
3. Experiment even more!
“However you think it should be, that’s exactly how it should be.”