A digital render of a country landscape with trees and mountains. At the centre is an abstract white otherworldly shape springing into the sky,
Somerset House Studios
Somerset House Studios

Click and Collect: Show me your dataset

PATH AI: Privacy, Agency and Trust

Sat 18 Feb 2023
10.30 - 17.30
Lancaster Rooms & River Rooms
New Wing

Artists, curators, researchers and storytellers reflect upon the current ecology of AI through discussion, performance, a workshop and a tour through the artwork from the PATH AI Residency Programme. 

Who decides what the future will look, feel or sound like? When thinking about AI conversation tends towards its biases and asymmetric development and implementation. It is clear that the goals behind its mainstream development haven’t been human-centered nor have they had a horizontal/diverse set of voices embedded in its design. Science fiction has played a very interesting role in the way in which both specialized and non-specialized agents have approached AI. 

The strong narratives behind ideas such as singularity and AI overpowering humans have been written and rewritten in novels, movies, podcasts, videogames and more, but how do these narratives impact our relationship with AI? Can we build new narratives which include a more horizontal, sustainable and caring AI? How can we speculate about the future of AI by rethinking what privacy, agency and trust mean for these systems? 

To dive deeper into these questions, Somerset House presents Click and Collect: Show me your dataset, a one-day event where a group of artists, curators, researchers, and storytellers reflect upon the current ecology of AI. This event includes discussions, talks, a performance, a workshop, and a commented tour through the resulting artwork from the PATH AI Residency Programme.

Curated by Doreen A. Ríos



11.00 - 11.10 | Introduction

Presentation by Doreen A. Ríos with AI-host, Iztli

11.10 - 12.10 | On human-centered design for AI systems | Panel Discussion

This panel focuses on how to think about AI from a human-centered perspective. It questions whether it is possible, or useful, to consider the sensible and the sublime as essential elements for developing an AI and whether or not this can be used to build a less biased dataset. Looking deeper into the work of Daniel Chávez Heras and his research about moving images, computers and qualitative data as well as Kristina Pulejkova’s work about memory creation as a starting point for the emergence of sentient AI, we’ll explore what it means to bring the human experience into AI systems. Moderated by co-founder of Feminist Internet, Dr. Charlotte Webb, this panel will be asking tough questions regarding the politics of AI, its apparent “neutrality”, and its current ubiquitous presence.

Moderated by Charlotte Webb
Participants: Daniel Chávez Heras and Kristina Pulejkova

12.15 | Prompt Battle | Talk (online)

In 2022, the development of machine learning models capable of transforming text prompts into images made a quantum leap in both quality and accessibility. DALL·E 2 seems to be ahead of the curve, but as the runners-up demonstrate, a competitive edge can be gained, not just through higher degrees of technical precision (Imagen), but also through a more permissive, meme-friendly, policy when it comes to content moderation (Craiyon), or by granting its users full ownership over their creations (Midjourney). In this talk, Florian A. Schmidt and Sebastian Schmieg share their insights and process behind the development of Prompt Battle, a live event they developed in collaboration with design students at HTW Dresden where people compete against each other using text-to-image software. 

Speakers: Florian A. Schmidt and Sebastian Schmieg

13.10 | Lunch

14.00 - 15.00 | Platform Specters: Deliberate Images, Intended Metaphors | Talk (online)

This talk reflects upon the politics of on-demand scalable platform labor with its focus on one of the most notorious online marketplaces to exist today. The platform, Amazon’s Mechanical Turk (MTurk), has a specific history that is embedded as a regulated non-presence at the level of the metaphor (the magic of technology; the performing ghost; following the metaphor of the cloud, you cannot see it by looking at it) and at the level of the real (people spread across the world perform digital piecework, concealing human labor in a manner that parallels an 18th century automaton declared to have achieved ‘total automation’). This history, as articulated by Mary Gray, started with slavery to ‘voluntary’ work, piece work, the Fordist assembly line, dot-com outsourcing and, finally, to online digital assembly lines.

Speaker: Aarti Sunder

15.10 - 16.10 | Speculative narratives for building ecosystems of care | Panel Discussion

This panel focuses on the ways in which storytelling and the development of narratives can help us break the mold of our relationship with several technologies, including AI. It will explore the work of the collective Interspecifics and their research regarding speculative communications, the collaboration with non-human intelligences and the creation of ontological machines. We’ll dive into questions around protocols, radical collectiveness, and the role of care within narratives that look into the human experience yet expand their reach into other living organisms. Moderated by artist and researcher, Maro Pebo, this panel will reflect upon questions such as how do these narratives impact our relationships with AI? Can we build new narratives which include a more horizontal, sustainable and caring AI? How can we speculate about the future of AI by rethinking what privacy, agency and trust mean for these systems?

Moderated by Maro Pebo 
Speakers: Interspecifics (online)

16.15 | Break

16.30 - 17.00 | Vica Pacheco presents ANIMACY | Performance

With her sound sculptures, Vica Pacheco, creates a sonic environment in formation. Animacy finds in the vessel the possibility of the empathy of the one who shares its own physiology to contain, but also to liberate and summon. Pre-Columbian technology mirrors contemporary technology to explore communication in its multiplicity of directions and rhizomes that go beyond the body. Animism does not limit itself to going back or forward in time in order to be able to update and mimic itself into a bird, a jaguar, a sunset or a deathbed. It never completes its metamorphosis and discovers, by creating extensions, a point of overflow for the limits of the body. The vessels that make up this orchestra are a reflection of the communication technologies to which Pre-Columbian ancestors entrusted their future.

River Rooms

10.30 - 17.30 | Random materialized library

A selection of AI papers, essays and stories. A reading area where you can get a random text printed for you to take home, read on the spot and/or exchange with someone else. These texts were shared by our collaborators and represent an essential part of their our perspectives around AI.

10.30 - 17.30 | PATH AI Residency Programme | Exhibition / screening

Somerset House Studios and the UAL Creative Computing Institute, in partnership with The Alan Turing Institute, the University of Edinburgh, and the RIKEN research institute in Japan, launched the PATH-AI residency programme. Commissioning three artists to create new works critically engaging with intercultural ideas of privacy, agency, and trust in relation to artificial intelligence (AI) and other data-driven technologies, the programme aimed at artists interested in exploring the international landscape for AI ethics, governance, and regulation. These are the results of the 6-months residency programme.

Ana Min Mein (Where am I From?), 2022,  Nouf Aljowaysir (SA/US) 
Cloud Song, 2022, Chris Zhongtian Yuan (CN/UK)
Los caídos, 2022, Juan Covelli (CO)

14.00 - 15.00 | Commented tour with artists | Talk 

Artists Nouf Aljowaysir and Chris Zhongtian Yuan will guide us through the artworks the created for the PATH-AI residency programme, focusing on their personal reflections upon the current state of AI. They wil share their processes as well as their inspirations and the ideas behind Ana Min Mein (Where am I From?) and Cloud Song.

Speakers: Nouf Aljowaysir and Chris Zhongtian Yuan


11.10 - 13.10 | Show me your dataset: text to image models | Workshop

The data (words, images, inputs) that go into machine learning models is central to the eventual output. This workshop will explore the content of several canonical datasets that are frequently used for computer vision, as well as trying to discover what has gone into training recent large language models (obscured and opaque). Participants will start to build their own datasets, concentrating on words that are difficult to define, images that are hard to classify and languages that no longer exist.

Led by Anna Ridler

14.00 - 16.30 | A·kin: Aarati Akkapeddi | Workshop

Artist, coder and educator Aarati Akkapeddi will be running a workshop that looks at Tamil studio photographs and different ways to read and organise them. Participants are encouraged to bring personal photographs, especially those taken in Tamil photography studios, to be collectively grouped with photographs from the stars.archive (Studies in Tamil Studio Archives and Society). Akkapeddi will share different approaches to image categorisation using both computer programmes and more personal points of view. The artist will then show how to digitally average similar images into a single composite image.
The workshop builds on Aarati’s current exhibition A·kin at The Photographers’ Gallery, London. Through the stars.archive and Aarati’s own family photographs, the exhibition asks what computers ‘see’ when looking at family albums. Are algorithmic models able to grasp the cultural traces of a community when looking at their archive of photographs?
This workshop is presented in partnership with The Photographer’s Gallery. Participants need to bring a laptop and have a Google account. A couple of laptops are available for those without access to one. Please email digital.programme@tpg.org.uk to check availability.


Privacy, Agency, and Trust in Human-AI Ecosystems (PATH-AI) is a collaborative and multidisciplinary research project between The Alan Turing Institute, the University of Edinburgh, and the RIKEN research institute in Japan. The aim of the project is to examine how the three interrelated values of privacy, agency, and trust work together in the very different cultural contexts of the UK and Japan in relation to AI and other data-driven technologies. Grounded in new research summarised in the PATH-AI interim report, the residency supports artists to engage with the report’s themes and findings, from the theoretical and cultural background of these values, to how they are being expressed and performed through everyday lived experience during the COVID-19 pandemic, with the aim of exploring how differing intercultural understandings of these values can inform the ongoing shaping of the international landscape. This work is supported by the Economic and Social Research Council and the Japan Science and Technology Agency.

With thanks to