This is an ongoing list of algorithmic collective action cases. We encourage you to contribute to it by creating a markdown file with the case and opening a pull request.
For more background and related work see Sigg et al. (2024).
Note that some of the strategies described here no longer work, e.g. due to changes in the algorithm. For many strategies it is unclear how successful and effective they are or were or how many people participated. Lastly, some strategies described here are more like prototypes or proof-of-concepts, but they nevertheless provide value by their ideas.
We group the cases by the mechanism they apply as a lever. This can be:
-
Collective information sharing
Digital platforms are often characterized by a high degree of information asymmetry between platform participants and platform owners. Workers have started to engage in information sharing strategies of different kind to counter this information balance. Examples include building up knowledge how the algorithm works, exchanging individual strategies to make them available to a larger group of users, or aggregating data as part of a data cooperative. -
Collective strategizing
This refers to cases where people agree upon and execute a joint strategy which often requires coordination as it might happen at a dedicated time or place. This can be viewed as a collective strategy during deployment time as it leverages certain aspects of the algorithm when deployed.
Moreover, the reliance of machine learning powered platforms on human-generated data during training time offers an additional lever for workers to exploit, e.g. by manipulating, poisoning or withholding training data.
| case ↓ | short case description | participants | area | category | benefit for participants | coordination aspects (if known) |
|---|---|---|---|---|---|---|
| Amazon Mechanical Turk – TurkerView & Co. | collecting/averaging ratings on requester, sharing individual strategies on how to accomplish tasks | workers | gig economy | collective information sharing | more accurate information about tasks and requesters | forum, websites |
| Uber Surge Verification Groups | drivers notify each other to verify accuracy and profitability of surge areas | workers | gig economy | collective information sharing | information can save each other time and money | messenger |
| John Deere Tractorhacking | collective effort to hack and share tractor software | tractor-owner, software-hacker, mechanics | collective information sharing | reduced costs | open source code platforms, e.g. GitHub | |
| Truck Drivers Facing Surveillance | electronic logging devices: brute-force destruction, exchanging of individual strategies, sharing log-in information, slow rolls, collaboration between drivers and truck stops | mainly workers | collective information sharing , collective strategizing | evading surveillance | ||
| Instagram Joint Sensemaking | influencers collectively decipher the logics of the algorithm by reporting which experiments worked out | content-providers | social media | collective information sharing | better/faster knowledge about the Instagram algorithm | messenger |
| Upwork Joint Sensemaking | upwork freelancers using forums etc. to jointly making sense of the algorithms logic on e.g. account-visibility, customer ratings etc. | workers | gig economy | collective information sharing | better/faster knowledge about algorithm | forum |
| MozillaRally | browser extension (no longer active) where participants could donate their data for web-based studies | browser users | collective information sharing | different from study to study | centrally organized, not user-driven | |
| Food Delivery in China – WeChat groups | drivers share information about individual strategies, trying to collectively unpack the platforms’ algorithm, messenger groups act as informal unions, sharing information on temporary subsidies, newcomer help etc. | workers | gig economy | collective information sharing | better/faster knowledge about algorithm, individual strategies | messenger invites via QR-codes on drivers' bags |
| Workers Info Exchange – Pooling Data | provide support for workers to make a data request from the gig platform, conducting research with that data | workers | gig economy | collective information sharing | expose unfair algorithmic decision making | |
| Driver’s Seat Cooperative | driver-owned data collection/analytics service, providing performance data for drivers, sell data to cities and transportation agencies | workers | gig economy | collective information sharing | performance data insights | app |
| WeClock | self-tracking app for workers, quantifying the workday; also addressing unions to support their campaigns | workers | gig economy | collective information sharing | performance data insights | app |
| Postmates BlitzUp | one-day protest as reaction to change in pricing policy; only accepting "Blitz" offers (higher paying orders) | workers | gig economy | collective strategizing | protesting and still earning money | campaign website with registration + action plan |
| Amazon Flex Treephones | drivers placing phones in trees + syncing them with their phones, gives advantage over location-dependent offers | workers | gig economy | collective strategizing | getting offers before others do without the need to be physically close to the store | potentially onsite |
| Uber Surge Club | collective log-out to trigger supply- and demand-based surge pricing, logging back after a few minutes yields higher pay | workers | gig economy | collective strategizing | getting higher payment for a ride | messenger |
| Youtube – Fix Community Poll Feature | campaign to extensively use community polls (which apparently reached users beyond measure) in order to 'fix the broken feature' | content-providers | social media | collective strategizing | getting fair conditions for everybody | video |
| Spotify – 31 Seconds of Silence | the band 'Vulfpeck' made money by asking their fans to listen to songs that consist of 31 seconds of silence during nighttime | users | collective strategizing | band gets money, fans get concerts financed by that money | ||
| Spotify – Song Ranking | singer 'Anitta' pushed her song to the first place on Spotify charts by asking her fans to listen to it repeatedly, create playlists and use multiple accounts | users | collective strategizing | fans could win a spotify subscription, singer gets fame + money | ||
| DoorDash – #DeclineNow | drivers reject low-paying orders, hypothesizing that the algorithm will increase pay on that order for fellow driver | workers | gig economy | collective strategizing | increased income | facebook groups |
| FoggySight – Facial Recognition Privacy | enhancing user privacy in facial lookup settings by producing ambiguous images to obfuscate another user’s identity | users (e.g. of social media) | social media | collective strategizing | protecting (other) users' identities | |
| Instacart – 22 cent Tips | tipping 22 cent upfront as a signal that more tips will be paid later | workers, consumers/clients | gig economy | collective strategizing | workers: more money, consumers: tipping the driver (not Instacart)" | campaign website, flyers |
| Instacart – #DeleteInstacart | workers asked customers to delete the app | workers, consumers/clients | gig economy | collective strategizing | unclear | hashtag, open letter |
| Food Delivery in China – Restaurant-Driver Cooperation | drivers cooperating with restaurants, placing artificial orders to meet bonus pay requirements | workers, restaurants | gig economy | collective strategizing | using the bonus-feature of the app to make more money | workers: money, restaurants: unclear, probably visibility |
| Instagram Engagement Pods | peer community to mutually engage with each others content | content-providers | social media | collective strategizing | improve their ranking or account visibility | |
| Instagram Loop Giveaways | multiple accounts collaborate to offer a gift for their followers, requiring participants to follow all collaborating accounts | content-providers | social media | collective strategizing | increase followers, improve ranking or account visibility | private messenger groups |
| Upwork – Evaluation Metrics | workers are asking clients to give them specific ratings to enhance their scores | workers, clients | gig economy | collective strategizing | better job score | |
| Upwork – Off-Platform Work | workers use the platform for matching with customers, then suggest to continue off-platform to earn more | workers, clients | Gig Economy | collective strategizing | private messages | |
| Twitter – Screenshotting | posting screenshots of other posts to avoid a forced connection | content-providers | social media | collective strategizing | not getting unwanted attention, avoiding forced connection | common knowledge, hints and comments about the strategy on Twitter itself |
| Twitter – Voldemorting | not mentioning words or names in a post (instead inventing new names) in order to avoid a forced connection | content-providers (+ content-consumers) | social media | collective strategizing | not getting attention, avoiding forced connection | |
| Facebook – Remote Geotagging | collective remote check-in to obfuscate geolocation data allegedly monitored by police | Facebook users | social media | collective strategizing | hide the true location data | |
| Google Bombs | heavy linking of search terms and websites to rank them higher | users | collective strategizing | push the website's ranking | ||
| Waze – Traffic Routing | pseudo-collective: individuals creating bots to influence crowd-sourced navigation systems | pseudo-collective data poisoning | ||||
| Fan Fiction Writers – Irreverent Stories | fan fiction writers produce fake stories to poison data collection services | workers | collective strategizing | |||
| AdNauseam | browser extension to obfuscate browsing and ad data aiming at protecting own privacy, protecting other users' privacy as a side effect | browser users | data poisoning/obfuscation | hide true user profiles | ||
| Impact app | prototype app to organize supporters to promote political messages on social media | social media users | social media | collective strategizing | steer attention on social media | app |