CROSSLUCID Art Summary

Video

Watch presentation at StreamETH

Summary

CROSSLUCID Art is an artistic practice exploring the creative possibilities of generative AI models. They have worked on generating diverse anthropomorphic representations by carefully constraining model parameters, as well as using hypnotherapy-inspired techniques to reshape latent spaces and alter violent/disturbing content. One key project involved creating a multi-channel, evolving narrative installation by combining text and visuals from different models.

Their current focus is developing an immersive, data-rich experience around marine protected areas. Participants will navigate an AI-generated underwater environment using haptic/liquid interfaces, with inputs from real-world data, conversational agents, and inscribed “memories”. A key aim is to raise awareness of these regenerative ocean sanctuaries.

The artists are exploring collaborative models to steward and sustain such complex, evolving AI artworks long-term. Their work has been exhibited in galleries and commissioned for publications on the future of AI art and ecosystems.

Key Takeaways

• Constraining generative AI model parameters can produce more diverse, unbiased outputs
• Exploring latent spaces allows altering undesirable content like violence
• Multi-modal experiences combining visuals, haptics, and language offer rich artistic potential
• Immersive data-driven AI environments can raise awareness of complex issues
• Evolving AI artworks require new sustainability/governance models beyond individual creators
• Collaborating with institutions provides resources and reach for impactful AI art projects
• Generative AI enables novel forms of storytelling and world-building
• Responsible development is key for positive societal impact of generative AI systems

Speakers

• Speaker A (No name provided)

• Role: Artist with CROSSLUCID Art

• Expertise: Generative AI, visual art, immersive experiences

• Contributions: Insights into their creative process, technical details of projects

• Speaker B (No name provided)

• Role: Artist with CROSSLUCID Art

• Expertise: Photography, film, architecture, generative AI

• Contributions: Background context, project overviews, collaborative models