Whether your work is in community development, well-being, or Artificial Intelligence (AI), the link between these fields might not be immediately clear. The goal of this short essay is to provide a perspective for why the work on well-being is critically important for the field of AI and therefore the need to bridge the gap between these epistemological frames.
As I type these characters on my clunky device, as you read these words, without a doubt I think we will both agree on one thing - time is passing. How we spend our time has a remarkable impact on our lives and the lives of everyone around us. Sitting in a Deep Learning class at Stanford, CA in 2016, I remember the prominent AI professor Andrew Ng speaking to Time as our most important resource when developing an AI algorithm. Deep Learning is a branch of AI which has to do with pattern recognition - the ability of a computer program to learn to recognize patterns in large amounts of data. Deep Learning algorithms have given rise to some of the biggest success stories in AI through solving problems related to image recognition, language generation and processing, and others.
Back in 2016, I was a working as a Senior Machine Learning engineer at an innovation lab in the Bay Area and had spent several years trying to solve complex problems through developing AI algorithms. I was constantly learning through trial and error and actively challenging the traditional approach to algorithm design which I was taught in University by questioning what is the best use of our time as AI engineers. To me, the answer to that question has always been related to our ability to envision the end goal and multidimensional impacts of the AI algorithms we are working on.
Oftentimes engineers are part of large and highly specialized teams which work on specific components of complex technologies. Existing organizational culture and structure incentivize speed and complexity which make it challenging and often impossible for an individual engineer to understand the larger so-called socio-technical context of their work - the social implications of the technical work they are doing. Asking the question “What is the best use of my time?” invites us to question our motivation for the work we are doing. For example, some of the challenges in designing AI algorithms stem from the business need to create systems that work at scale. Traditionally, the main metric for success is the so called ability of an AI model to generalize i.e. to perform the task it was developed for in new contexts which differ from the development context. To solve for that, engineers often need to make tradeoffs in deciding what experiments will allow them to learn the most about how the AI algorithm they are developing will behave in unseen and dynamic real world contexts. For example, should they collect more data on which to test their algorithm, what kind of data should they collect, should they modify the parameters of the AI algorithm, what algorithm should they try next, etc. It is often easier to solve a problem in a context we are familiar with. This, however, may lead us to forget the complexity of the real world impact of the systems we develop and instead see that impact reduced to numerical AI optimization errors we are trying to minimize.
By keeping present our deep motivation for the work we are doing, we have an opportunity to see how AI changes the fundamental nature of human experience. Seeing ourselves as participants in that change helps us realize our responsibility for the visible and invisible ways in which the work we do impacts the well-being of people and other beings. Understanding AI impacts is perhaps beyond the cognitive capacity of any single individual and therefore invites us to seek collaborations especially with people from interdisciplinary fields. Adopting a well-being lens in the work on AI holds a promise to bring more clarity, integrity, and response-ability that could enable us to collaboratively design systems that are better aligned with our individual and broader societal values.
Guest post by Bogdana Rakova, Member of the Board of Directors of the Happiness Alliance
Read more about AI and Well-being in the Special Issue: Intersections of Artificial Intelligence and Community Well-Being
No comments:
Post a Comment