Reddit bot experiment shows users resist simple AI manipulation
Bots giving random awards actually reduce user activity on Reddit, study finds.
A field experiment on Reddit tested whether users respond to symbolic awards from apparent human and bot accounts. Awards were given with four rationales: praising logical argument, emotional sensitivity, moral integrity, or explaining the award came from a random lottery. The awards did not increase user activity or downstream impact, and awards from bots with the lottery rationale actually reduced them. However, awards encouraged direct communication between users. These findings suggest online users can be resilient to simple behavioral manipulation from platforms and artificial agents, but not necessarily to more sophisticated schemes that mimic human conversation. Transparent labeling of automated agents remains essential.
- Awards from bots using a 'lottery' rationale reduced user activity and downstream impact on Reddit.
- Overall awards did not increase user engagement but did encourage direct communication between users.
- Transparent labeling of automated agents is essential for ethical platform governance, per the study.
Why It Matters
Demonstrates users' resilience to simple AI manipulation but warns of sophisticated deception schemes.