Antidepress Bots
Project details
- Year
- 2021
- Programme
- advertising-beyond
- Practices
- Autonomous
- Minor
- Critical Studies
Antidepress Bots is an activistic system of social media bots that distorts Instagram’s algorithm’s ability to push depressive content to depressed people.
The bots were exhibited so their rebellion against Instagram’s algorithm could be viewed live.
Just before the start of the COVID-19 lockdowns I came out of a depression during which I noticed that social media algorithms played a role in amplifying my negative thoughts. Recommendation algorithms curate what posts we see when we log onto social media. The algorithms observe what content people interact with to learn what we like. However, the algorithms can’t tell what content is good or bad for us. Many depressed people interact with depressive content, but as a result the algorithm learns to push this negative content to their feed. These users are placed in the ‘depressive filter bubble’; a digital space where their depression is amplified.
By programming bots (which are automated instagram accounts behaving like real users) to follow only accounts that share depressive content, the algorithm places them in this depressive filter bubble alongside actual depressed people. Then, from within that bubble, the Antidepress Bots are programmed to only interact with images of clear blue skies.
This confuses the algorithm into pushing blue skies to depressed people in- stead of gloomy content. Clear blue skies aren’t only visual antidotes to the dark and depressive content, but blue light has also been implemented as an effective suicide prevention method on train stations.
More running Antidepress Bots will mean a greater effect of positive distortion. This is why the code of the project is open-source; anyone is encouraged to download and run the bots to help relieve depressed social media users world-wide. It is a collective activist stance against the algorithms that negatively influence vulnerable people.
Technical side of the project actualised in collaboration with Jasper Kamphuis and Robbert Koning.