Aligning AI Optimization to Community Well-Being
- PDF / 505,873 Bytes
- 21 Pages / 439.37 x 666.142 pts Page_size
- 51 Downloads / 218 Views
Aligning AI Optimization to Community Well-Being Jonathan Stray 1 Received: 7 February 2020 / Accepted: 9 October 2020/ # Springer Nature Switzerland AG 2020
Abstract This paper investigates incorporating community well-being metrics into the objectives of optimization algorithms and the teams that build them. It documents two cases where a large platform appears to have modified their system to this end. Facebook incorporated “well-being” metrics in 2017, while YouTube began integrating “user satisfaction” metrics around 2015. Metrics tied to community well-being outcomes could also be used in many other systems, such as a news recommendation system that tries to increase exposure to diverse views, or a product recommendation system that opstimizes for the carbon footprint of purchased products. Generalizing from these examples and incorporating insights from participatory design and AI governance leads to a proposed process for integrating community well-being into commercial AI systems: identify and involve the affected community, choose a useful metric, use this metric as a managerial performance measure and/or an algorithmic objective, and evaluate and adapt to outcomes. Important open questions include the best approach to community participation and the uncertain business effects of this process. Keywords Artificial intelligence . AI ethics . Community well-being . Optimization .
Corporate social responsibility
Introduction This paper is an extended analysis of a simple idea: large-scale commercial optimizing systems may be able to manage harmful side effects on communities by monitoring established well-being metrics. It sketches a theory that ties together quantitative measures of well-being, contemporary metrics-driven management practice, the objective function of optimization algorithms, participatory and multi-stakeholder governance of algorithmic systems, and the protection or promotion of community wellbeing. Detailed analyses of recent efforts by Facebook and YouTube are used to
* Jonathan Stray [email protected]
1
Partnership on AI, San Francisco, CA, USA
International Journal of Community Well-Being
illustrate the challenges and unknowns of this approach, which generalizes to a variety of different types of artificial intelligence (AI) systems. The core contribution of this article is a proposed process for the use of community well-being metrics within commercial AI systems. Well-being encompasses “people’s living conditions and quality of life today (current well-being), as well as the resources that will help to sustain people’s well-being over time (natural, economic, human and social capital)” (OECD 2019b, p. 2). Community well-being attempts to evaluate well-being at the level of a community defined “in geographic terms, such as a neighborhood or town … or in social terms, such as a group of people sharing common chat rooms on the Internet, a national professional association or a labor union” (Phillips and Pittman 2015, p. 3). The measurement of well-being is now a well-
Data Loading...
