1,548 days, 2,379 entries ... Newsticker, link list, time machine: HOLO.mg/stream logs emerging trajectories in art, science, technology, and culture––every day
Year 2023 2022 2021 2020 Show All
Month January February March April May June July August September October November December Show All
OUT NOW :
Wendy Hui Kyong Chun
Discriminating Data
The
media theorist and systems design engineer reveals how big data and machine learning encode discrimination and create agitated clusters of comforting rage
“The algorithm estimates Hito’s gender, it says she is 57% female and 42% male. Which begs the question: what would 100% female be? Whether that’s Barbie, Grace Jones, or Angela Merkel—who knows?”
– Trevor Paglen on
Machine-Readable Hito (2017), which tasked facial analysis algorithms with guessing Hito Steyerl’s age, gender, and emotional state across hundreds of photos. In dialogue for SJMA’s “
Artists in Conversation ” series, the duo discuss representation, truth, and power relations.
Artist and composer Holly Herndon launches Holly+ , her own AI twin that will interpret any polyphonic audio uploaded to holly.plus . Problematizing voice ownership and artist compensation, Holly+ and subsequent AI voice tools are owned by a DAO cooperative, and any income generated will go toward new developments. “Vocal deepfakes are here to stay,” states Herndon. “A balance needs to be found between protecting artists, and encouraging people to experiment with a new and exciting technology.”
Seoul-based light artists Kimchi and Chips install Halo at the local National Museum of Modern and Contemporary Art (MMCA) as part of the institution’s “Multiverse” marathon. In the piece, 99 robotic mirrors follow the Sun’s movement and focus its light into a glistening halo within a cloud of fine water mist. First shown at London’s Somerset House in 2018, the MMCA edition features a new ‘solar farm’ system that uses machine learning for tracking and calibration.
“The closer the research started getting to search and ads, the more resistance there was. Those are the oldest and most entrenched organizations with the most power.”
– A Google employee with experience of the company’s research review process on the rejection of
Timnit Gebru ’s critique of natural language models (“
On the Dangers of Stochastic Parrots ”) that ultimately led to her firing. Writer
Tom Simonite traces the career of the former Google Ethical AI researcher and pieces together what
really happened when the company forced her out.
OUT NOW :
Slanted #37
Artificial Intelligence
An extensive survey of how AI and machine learning are changing art, design, and society at large, featuring the work of
Jon McCormack ,
Sofia Crespo ,
Douglas Coupland , and many others
“I for one enjoy watching Mercedes Benz simulate the Autobahn. It’s the Kraftwerk in me.”
–
Peter Kirn , musician and creative technologist, on NVIDIA’s
GTC 2021 keynote , in which the GPU giant demonstrated how, for example, car manufacturers are training AIs for future robotic factories in the company’s simulation platform
Omniverse . On his blog
CDM , Kirn explains why these advanced enterprise applications are relevant for artists, too
“The reason we use this metaphor of a baby is Spawn only has access to the information that we give her—we foster her with the data we feed her. We found this a way to not only talk about the importance of her daily diet, but the communal nature of raising a nascent intelligence.”
In the second of a three-post series revealing their spatial computing roadmap, Facebook Researchers reveal in-progress gestural and neural interfaces. One demo shows expressive wrist and finger control mediated by a watch-like wearable reminiscent of the Myo armband ; supporting concept videos telegraph aspirations for neural keyboards and that old AR chestnut: a user manipulating GUI elements in 3D space in front of them. However far off the tech is, it’s immediately satisfying to watch one of the researchers gush about the shiny future of neural interfaces where “you and the machine are in agreement about which neurons mean left and which ones mean right” without any mention of data harvesting or his employer’s long history of malfeasance.
With their solo show “In the Absence of Evidence,” opening at Galerie Liusa Wang, Paris, artist duo Quadrature revisit the boundary between science and fiction: C.R.E.D.O. (Cosmic Radio Engine for Delusional Observations), for example, is a 2020 sound installation that channels Pulsar signals, received via an on-site radio dish, into an otherworldly score. “The raw signals serve as input data for a neural network trained on human theories on aliens,” state the artists. The composition is the network trying to find E.T. messages in the noise.
“We refer to these attacks as typographic attacks . They’re far from simply an academic concern.”
–
OpenAI researchers, on the ‘blindspots’ of the lab’s latest computer vision model
CLIP . While CLIP shows remarkable capacity for abstraction—its multimodal neurons respond to literal, symbolic, and conceptual representations—it is also easily fooled: “when we put a label saying ‘iPod’ on this Granny Smith apple, the model erroneously classifies it as an iPod.”
“There is the sense of witnessing a private, secretive, and interior process, yet one contrasted by the vast exterior of open-source images on which it feeds.”
– Critic
Forrest Muelrath , on the “disturbing, hallucinatory” scenes on view at Casey Reas and Jan St. Werner’s
bitforms gallery exhibition. “Alchemical” builds on Reas’
Compressed Cinema series, in which a carefully trained GAN model reimagines a corpus of feature films. “It’s not quite dream imagery,” writes Muelrath, “more like unconscious thoughts before they can be articulated.”
“In my mind, for the last few years, the hand axe had been a stand-in for technology influencing human cognition. So when GPT-3 came up with that image, I was like ‘wow!’”
–
K Allado-McDowell , in conversation with
Holly Herndon and
Mat Dryhurst , on co-writing the book
Pharmako-AI (2020) with OpenAI’s autoregressive language model. “Not only were my ideas being continued and added to in seemingly novel ways—it almost felt like the AI reflected things I was subconsciously putting in the text.“
Artist-researchers Adam Harvey and Jules LaPlace , in collaboration with the Surveillance Technology Oversight Project (S.T.O.P.), launch Exposing.AI , an online tool that lets users find out whether their Flickr photos have been used for training commercial face recognition and biometric analysis systems. The web app scans across twelve notorious datasets and provides deep analysis: MegaFace (image), for example, includes 3,311,471 Flickr photos used by Amazon, Google, and other corporate giants.
American digital artist Casey Reas and German composer Jan St. Werner (of Mouse on Mars fame) get “Alchemical” at New York’s bitforms gallery. The show presents Reas’ experiments with machine learning techniques as image-making instruments in which he mobilized generative adversarial networks (GANs) trained on feature films. On view: a series of haunting Untitled Film Stills (2021) and five Compressed Cinema films that Werner augmented with “sentient” electroacoustic sounds.
“I thought I would have at least 10 years left to prepare for the future where everyone has their instant anything digital on demand at home. Looks to me like it’s down to something like five years now.”
– German AI artist
Mario Klingemann , on the arrival of OpenAI’s
DALL-E , a powerful neural network that creates images from text. “The only thing that prevents the internet from overflowing in strange new memes is that there is no open access to this yet,” writes Klingemann.
“By channelling patterns of fluid single-cell forms and the spiritual presence of artificial intelligence, she opens up the possibility of an encounter with other beings who have been ‘in the room’ all along: our cohabitants, symbiotes and other ghosts in the machine.”
–
Gary Zhexi Zhang , on the work of Finish artist
Jenna Sutela who “incorporates shapeshifting slime molds, algae blooms, and machine learning algorithms into her practice to foster a new understanding of the world”
“Uncanny Valley: Being Human in the Age of AI” opens at San Francisco’s de Young Museum, proposing “new ways of thinking about intelligence, nature, and artifice.” In showing works by Zach Blas , Ian Cheng , Simon Denny , Stephanie Dinkins , Trevor Paglen (image), Hito Steyerl , Martine Syms , and others, curator Claudia Schmuckli brings together essential artistic positions that parse the strange and unsettling intricacies of human-machine relations.
Load More
To dive deeper into Stream, please
Log-In or become a
HOLO Reader .
Daily discoveries at the nexus of art, science, technology, and culture: Get full access by becoming a HOLO Reader !
Perspective : research, long-form analysis, and critical commentaryEncounters : in-depth artist profiles and studio visits of pioneers and key innovatorsStream : a timeline and news archive with 1,200+ entries and countingEdition : HOLO’s annual collector’s edition that captures the calendar year in print
Become a HOLO Reader