Data & Society Workshop: Environmental Impact of Data-Driven Technologies
On November 2, 2018, Data & Society will host a workshop in NYC on the environmental impact of data-driven technologies. The purpose of the D&S Workshop series is to enable deep dives with a broad community of interdisciplinary researchers into topics at the core of Data & Society’s concerns.
Environmental Impact of Data-Driven TechnologiesBy the end of 2018, Bitcoin will consume .05% of the world’s energy per year. This is equivalent to the energy consumption of Denmark. Major tech companies are working hard to make cloud services more energy efficient, but server farms still require tremendous power and water to function. Additionally, other parts of the “stack” (e.g., software development, usage patterns) do not take environmental impact into consideration. Likewise, financiers obsessed with blockchain and 5G are often ignoring the environmental impact of the proliferation of these new technologies. While some IoT chipmakers are competing on energy efficiency, cheap production still dominates that conversation at a moment in which data-oriented tech is being introduced into everything.
On the user end, people are streaming a billion hours of YouTube videos every day and loading countless hours of videos and images into online backup services where they are likely to be watched/viewed by humans only a handful of times. Gmail has normalized the idea that everyone should archive email in perpetuity, which means that Facebook notices indicating you have a new message that you received in 2007 are still using up energy.
Apple has been called out for slowing down its operating system when battery life declines to make the user experience more seamless, which, in effect, encourages users to buy more equipment. Yet, the environmental cost of new hardware is piling up – quite literally. Users of Amazon Web Services and Microsoft’s Azure are encouraged to spin up new machines when they are working with data; they experience no visceral understanding of the environmental impact of their decisions. Likewise, even though most older computer scientists obsessed over runtime efficiency of their algorithms, few who grab code from Github give much thought to the environmental cost of their inefficient code.
Much work is still needed to understand the environmental cost of technology. The purpose of this workshop is to bring together researchers who are examining these issues from different disciplinary and analytic perspectives. Relevant topics for this workshop might include:
What is the environmental cost of blockchain, 5G, AI, and other hyped technologies?
How do design concerns at different parts of the “stack” affect the environmental impact of whole systems?
What would an environmental audit of artificial intelligence look like?
How do/might software engineers or other practitioners integrate climate concerns into their practice?
What is the relationship between privacy and energy-sensitive code?
How do data centers affect water policies in different countries?
How can decentralized engineering practices be made more environmentally responsible?
