Future in Flux How AI Is Changing Creative Workflows
summary
Explains that generative models often train on web data without consent and frequently hallucinate facts, showing early harms like auto-added artist signatures and MidJourney outputs that were uniformly white. Describes concrete tooling: Adobe Firefly trained on Adobe Stock with a contributor compensation model, in-app features like generative fill and generative expand in Photoshop, and content provenance via content credentials and the Content Authenticity Initiative. Notes persistent problems: bias, lack of consistency across scenes, heavy prompting, and an internal AI ethics review to mitigate harms.
AI-written summary · about
chapters (12)
- 0:00Introduction and speaker intro
- 1:54AI and machine learning basics
- 6:52Generative AI explained
- 9:01Ethical concerns with AI training
- 12:16Hallucinations and bias in AI
- 14:07Control and consistency issues
- 17:20Adobe Firefly and content licensing
- 22:12Historical context of technology fears
- 23:17Human vs machine creative strengths
- 27:16Future possibilities with AI
- 29:44Call to action for creators
- 30:12Q&A session
related talks
36mArt in the Age of Algorithms
40mThe Expanding Dark Forest and Generative AI
55mAI Showdown: The Good, The Glitchy, and The Risky
07m“I’m a Solo Developer but AI is My New Ill-Informed Co-Worker”: Envisioning and Designing Generat...
49mCreation as Curation [Keynote]
21mOpportunities of Generative AI
34mAI & Art: Redefining Creativity for a New Generation
28mEthical AI?!