PhD student at MIT, studying how data stewardship and analysis can impact community governance.
Right now, I'm focused on how algorithmic management is changing the reality of work, how data stewardship and participatory design can help create alternate working futures, and on the data rights of platform workers.
Recent Publications and Invited Talks
- FAccT2022Keynote: How to Bargain with a Black Box: Auditing an Algorithmic Pay Change With a Community-Led Audit
Willy Solis, Vanessa Bain, Dan Calacci, Drew Ambrogi, Danny Spitzberg
A real-world community audit of a black-box algorithmic system, the Shipt Calculator impacted workers, organizers and researchers and demonstrates how community-led research can be part of the FAccT community.
- CHIWORK2022Organizing in the End of Employment: Information Sharing, Data Stewardship, and Digital Workerism
Dan Calacci
Position paper in CHIWORK '22 arguing that a new "Digital Workerism" in the CHI and CSCW communities is needed to bolster the labor movement and balance information asymmetries.
- CSCW2022Bargaining With the Black-Box: Designing and Deploying Worker-Centric Tools to Audit Algorithmic Management
Dan Calacci, Alex (Sandy) Pentland
This paper introduces the Shopper Calculator, a tool used in a 2020 worker-led campaign that revealed that Shipt's new black-box payment algorithm cut the pay of over 40% of studied workers.
- Op-Ed2022Google Needs to Unlock Its Ad Privacy Black Box
Gizmodo
Google's FLoC was a proposal that would change the way the web fundamentally worked for millions of people. Why was studying it so inaccessible? In this Op-Ed, I argue that centralized gatekeeping of future web technologies is dangerous for the future of the web. I call for Google and other major companies to publish toolkits that let researchers study new technologies that will fundamentally change the web.
- Under Review2022Worker's Collective Data Access Rights: Adding Context to Worker Data Protection
Jake Stein, Dan Calacci (equal contribution)
How does current data protection law work for workers? This paper explores the limits of current data protection law in the US and EU for workers that seek to use their data collectively to build power.
- Under Review2022Privacy Limitations Of Interest-based Advertising On The Web: A Post-mortem Empirical Analysis Of Google's FLoC
Alex Berke, Dan Calacci
In 2020, Google introduced FLoC, a way to facilitate interest-based individual advertising without 3rd-party cookies. This paper shows that the FLoC proposal included serious privacy risks and explores FLoC's risk of leaking sensitive demographic information about it's users.
- CSCW2022The Cop in Your Neighbor's Doorbell
Dan Calacci, Jeffrey Shen, Alex (Sandy) Pentland
We use spatial regression models, structured topic models, and an experimental survey to understand how users of Amazon's Ring Neighbors network racialize and criminalize their subjects, and document what kinds of communities nationwide use the network most.
- Nature Comms.2021Mobility patterns are associated with experienced income segregation in large US cities
Esteban Moro, Dan Calacci, Xiaowen Dong, Alex (Sandy) Pentland
Is your local coffee shop more segregated by income than your favorite movie theater? We use a massive data set of mobile phone mobility to answer this question and model how individual segregation is related to people's tendency to explore new places and interact with those different than themselves.
- Data & Society2020Data & Society: Cop in Your Neighbor's Doorbell
Dan Calacci
Invited talk at Data & Society on mapping and analyzing Amazon Ring's network.
- HOPE2020One Ring to Surveil Them All: Hacking Amazon Ring to Map Neighborhood Surveillance
Dan Calacci
Remote presentation at HOPE (Hackers On Planet Earth) 2020 on hacking Amazon Ring's Neighbors app to reveal and measure the extent of neighborhood surveillance captured in the Ring Doorbell camera network.
- AAMAS2020Leveraging Communication Topologies Between Learning Agents in Deep Reinforcement Learning
Dhaval Adjodah, Dan Calacci, Abhimanyu Dubey, Anirudh Goyal, P.M. Krafft, Esteban Moro, Alex Pentland
Can network structures inspired by human social networks improve distributed reinforcement learning algorithms? This paper proves that arranging agents in different network topologies can massively improve evolutionary deep reinforcement learning algorithms.
- Op-Ed2020Location Tracking To Fight Coronavirus Is Dangerous And Possibly Pointless
Gizmodo
At the beginning of COVID, many states and universities were experimenting with using location data to track covid spread. In this op-ed I argued that location data is a dangerous technology to break out for state-level disease surveillance and is a poor technical choice for tracking airborne illness.
- Preprint2019The Tradeoff Between the Utility and Risk of Location Data and Implications for Public Good
Dan Calacci, Alex Berke, Kent Larson, Alex (Sandy) Pentland
Location data collected from mobile phones and aggregated in massive databases poses enormous risks to individual and collective privacy. It also poses clear utility for research, marketing, and policymaking. This paper explores and conceptually models the risks that large-scale location datasets introduce, and speculates on ways that location data can be regulated or protected while offering significant utility.