2019 - Critical AI Manifesto Application - The United Intelligence Lab (UIL)
Texto
THE COMMON MANIFESTO (texto del 20 de abril de 2022)
AI ethics/regulations should be based on the domain its being used for.
AI should adhere to feminist values, so it doesn't knowingly or unknowingly perpetuate gender inequality.
Humans should always be able to override AI.
AI's decision-making process should always be explained on request, live or in retrospect.
All code for AI should be open source and checked before release.
All datasets should be approved by a supranational (or transnatural) institution before the AI is released.
The outcomes (decisions) of AI's should be seen as opinions, not as facts.
Data always has context. Contexts change over time and thus also should the interpretation of data and the labels connected to it.
AI is not a mystic, impenetrable entity: it is created, trained and legitimised by humans. Not understanding an AI is not an excuse to not try to understand and explain how it works: not understanding should evoke curiosity to try to learn to understand.
AI laws should be designed by governmental organizations, not by corporate entities.
Contexto
Online Tool
V2_ is developing an online tool to position creatives in the vast and complex world of AI. The 'Critical AI Manifesto is a sounding board to scrutinise the technopolitical infrastructures of AI in order to spur more imaginative and critical dialogues regarding artificial intelligence’s impact and governance within the creative practices.
The online tool allows you to compose a personal manifesto by browsing through and selecting from a pre-existing set of statements that can be prioritized at your own discretion. MeetUp
This meetup is about sharing your perspective on AI and expanding the pool of statements by peer group discussion. Feel free to check out the beta version (https://caim.v2.nl/) in advance and if possible bring a laptop to the event.
https://v2.nl/events/the-critical-ai-manifesto
https://web.archive.org/web/20220421203720/https://v2.nl/events/the-critical-ai-manifesto
Más información sobre el grupo y la aplicación en
https://v2.nl/archive/articles/lets-wind-up-the-nursery-tale-of-ai
No hay una versión única ya que la aplicación está diseñada para que se puedan añadir afirmaciones "statements" continumente pero se consensua una versión común que es la que se incluye en el texto de arriba
Enlaces
Primera edición:
URL: https://caim.v2.nl/common-manifesto (aplicación)
Wayback Machine: https://web.archive.org/web/20220421203436/https://caim.v2.nl/common-manifesto