(Re)Defining A.I.

Colleen Estrada, Phil Van Allen

The (Re)Defining A.I. Working Group will discuss the purposes and meanings of machine intelligence, and how designers should approach it and influence it. We’ll question the very nature of intelligence, and if “artificial” intelligence is even the right way to think about this rapidly evolving field. We’ll be asking questions such as:

  • Should we be trying to create facsimiles of human intelligence or design new “alternative” intelligences?
  • When people collaborate with machines, what could happen to authorship (Links to an external site.), originality, authenticity and creativity? What do we want to happen?
  • What is A.I. as a design material? What is the grain of the material, and what are it’s affordances and limitations?
  • How can multiple, unique A.I. systems (Links to an external site.) change how people converse with A.I.?
  • What does Meaningful Control of A.I. (Links to an external site.) imply for designers? How autonomous should A.I. systems be?
  • Bias is inherent in A.I., so how do we work with data and behavior to mitigate the negative effects and help people (and designers) manage and understand bias?
  • How can AI be used in unexpected ways? I.e. what are the new, strange, beautiful, provocative uses of AI?
 
Alternate Description (this umbrella description was what I originally proposed last year when we were first talking about interest groups)
 
AI Ecologies
This interest group looks at speculative and applied approaches to Artificial Intelligence and the ecologies they create. We’re particularly interested in the complex behaviors and relationships created when “smart,” autonomous systems integrate into people’s lives. What are the inner lives of AI entities? What are their values, goals, intentions and where do those come from? How they communicate and interact with other AIs and with people? What does AI diversity mean and how does it affect the milieus they inhabit?
Within this interest group, there are several more specific areas of investigation:
  • Designed Animism — This research is developing a post human centered design approach to design, where smart things are given narratives as if they have agency, needs, goals, personalities, and moods rather than being tools or servants. In this ecology centered design framework, the goal is to create decentered, fertile, and productively serendipitous milieus where people and AI entities thrive and evolve.

    The research looks at new interactions, new design paradigms, and cultural shifts that come out of giving AI systems backstories and agency. We’re also interested in the challenges of designing interesting emergent systems that come out of a mix of people and diverse species of AI.
  • The Internet of Enlightened Things — This research explores the implications and opportunities of sharing our lives – willingly or not – with ever more “intelligent” objects and systems.  We are interested in new manifestations of Artificial Intelligence (AI) and Machine Learning (ML) in the neighborhood where people interact with the urban at a human scale.

    From intelligent street lights that track vehicles and pedestrians, to emotional recognition systems in retail stores; soon our urban environment could be full of autonomous AI systems that change the character (and constituents) of the “local.” What role should design play? What about the well being of the AI systems? What are the day-to-day neighborhood implications of the technology and methodologies of AI/ML – neural nets, supervised/unsupervised learning, the edge/fog/cloud network infrastructure, and the methods and biases of data-scientists?
  • Conversational Interactions — This work looks beyond chat bots, and digs into the broader realm of how people and machines communicate and collaborate. While voice interactions are important, we believe that other forms of communications should be a part of the mix. How can movement, sound, light and collaboration play out between people and AI, as well as between AI and AI? What are “native” expressions and behaviors of machines? How do we bridge the conceptual gaps between human and machine intelligence? What affordances open up when AI entities are “embodied” in virtual or augmented realities, and can actively collaborate with people and each other in creating new things?

More Projects