A profile of Holly Herndon, an artist and musician who uses AI in her work and who co-founded Spawning, a company building a “consent layer for AI” for artists
Anna Wiener / New Yorker : X: @newyorker , @kcesq , @totalvibration , @newyorker , @hollyherndon , @annawiener , @joelvmills , and @newyorker LinkedIn: Nicolas Smirnoff and Zoe Cooper X: @newyorker : @EyalPress @joshuarothman .@AnnaWiener spends time with Holly Herndon, an artist and musician who uses machine learning to make strange, playful pieces of work, and advocates for artists' autonomy in an A.I.-centric world https://nyer.cm/H27WMIt Kevin Casini / @kcesq : Fascinating read here by @hollyherndon who is uniquely qualified in this space to speak on such things - a paragon to look to as we venture forward in a world of AI art creation https://www.newyorker.com/... Lars Gotrich / @totalvibration : “What was more of a fork than a baby?” This @annawiener profile of @hollyherndon got me surprisingly in my feels, as a parent, as a creator, as a human. https://www.newyorker.com/... @newyorker : .@hollyherndon helped create a bespoke A.I. model that incorporates the voices of about 50,000 people. “We were trying to communicate that it's not just some kind of alien intelligence,” she said. “It's all of this human activity.” https://nyer.cm/3QwaB2h Holly Herndon / @hollyherndon : It has been a crazy few years (AI baby,AI,baby). Near death and new life. Thanks @annawiener for capturing things so tenderly in this @NewYorker profile. Thanks to everyone who spoke in support of our art! https://nyer.cm/QH9HA7l Anna Wiener / @annawiener : For The New Yorker's A.I. issue, I wrote about the artist and musician Holly Herndon; artists who use A.I.; and A.I. that uses artists: https://www.newyorker.com/... Joel Mills / @joelvmills : Fantastic, must read piece on @hollyherndon's creative work with generative AI, exploring what's possible creatively and what's at at stake on consent and authorship: “How do you build a new economy around this where people aren't totally fucked?” https://www.newyorker.com/... @newyorker : .@hollyherndon navigates the tensions between A.I. and art, extolling the potential of the technology while helping to protect human creators. https://nyer.cm/WxnDYAN LinkedIn: Nicolas Smirnoff : The The New Yorker profiles Holly Herndon in a brilliant piece titled “Holly Herndon's Infinite Art”. … Zoe Cooper : Happy to see Holly Herndon, one of our contributors at Impossible Object, profiled by Anna Wiener in The New Yorker's AI issue:
Context & Ripple Effects
Holly Herndon’s work places AI use and artist autonomy in the same frame: she makes machine-learning-based art while co-founding Spawning to pursue consent mechanisms for creators.
The profile arrives after [[a:984092|artists challenged the unconsented use of their work in a Stable Diffusion training dataset]], making consent a practical question for AI’s creative supply chain rather than only an artistic critique.
First-order effects
- The profile raises Spawning’s visibility as a company focused on a consent layer for artists, while positioning Herndon as both an AI practitioner and an advocate for creator control.
- For artists, it foregrounds consent as a condition that can be designed into AI workflows, rather than a concern raised only after training data has been used.
Second-order effects
- AI developers and creative platforms face greater pressure to show how they identify and honor creators’ permissions if they want access to cultural material.
- The debate extends beyond model outputs to the provenance of training inputs, alongside writers’ early experimentation with AI-assisted writing tools and novel generation.
Third-order effects
- If consent layers become broadly adopted, access to creative data could shift from largely unilateral collection toward permission, attribution, and potentially compensation systems.
- That shift would make the commercial value of creative AI depend not just on model capability, but on whether its input supply can be governed credibly.
The trend: Creative AI is moving toward a contest over who controls, authorizes, and captures value from the cultural data that models use.