Going beyond smart: Towards collectively intelligent environments
Johannes Castner
Ancoris, United Kingdom
: J Comput Eng Inf Technol
Abstract
If we truly want to improve our decision making, by basing it on data, we must go beyond current smart solutions. Smart cities, smart hospitals and smart businesses are meant to support smarter decision making. And while such systems surely have achieved some improvements in human experience--they’ve helped with decongestion of traffic, pollution and more. They cannot reliably help us make better decisions; for the simple reason that seeing the results of a system in the form of data is not the same as understanding how the data were produced. By and large, these smart systems don’t help us much with causal understanding. Importantly, they don’t help us at all in understanding how new decisions and innovations translate into outcomes. Our algorithms operate almost exclusively on data from the past. Often, in fact, such smart systems lead to impoverished experiences for humans and sometimes they seriously harm people. They rely on narrow optimizations of one, or a small number of metrics. The solution is to design smart solutions more collaboratively and to go beyond them by introducing collective intelligence systems where humans are in the loop. In particular, humans must be in the loop for information collection, theory development and they must be part of all optimizations.
Biography
Johannes Castner works in the data and AI practice at Ancoris, where he participates in ethical AI and business intelligence projects. Recent work includes smart hospital and smart city projects. He started his career at the Federal Reserve Bank of Boston where he built econometrics software and at eBay where he worked on complex NLP problems. Then, in 2018, Johannes founded CollectiWise where he gained expertise in modern deep learning, data pipelines, knowledge graphs and reasoning systems. CollectiWise builds a SaaS service that aims to leverage the beliefs of all of a collective’s members to ultimately generate collective intelligence.
